[clang] 91eda9c - [NFC] Recommit "Regenerate SVE ACLE intrinsics tests"

Usman Nadeem via cfe-commits cfe-commits at lists.llvm.org
Thu Sep 2 18:30:17 PDT 2021


Author: Usman Nadeem
Date: 2021-09-02T18:28:44-07:00
New Revision: 91eda9c30f33da6ec6da70b59a5f5da6c6397039

URL: https://github.com/llvm/llvm-project/commit/91eda9c30f33da6ec6da70b59a5f5da6c6397039
DIFF: https://github.com/llvm/llvm-project/commit/91eda9c30f33da6ec6da70b59a5f5da6c6397039.diff

LOG: [NFC] Recommit "Regenerate SVE ACLE intrinsics tests"

Change-Id: Ida45fc41231cd71709048f2d37f228f14053514e

Added: 
    

Modified: 
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abs.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acge.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acgt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acle.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_aclt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_add.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adda.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_addv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_and.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_andv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asrd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfdot.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmmla.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bic.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brka.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkn.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpa.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cadd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cls.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clz.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmla.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpeq.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpge.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpgt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmple.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmplt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpne.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpuo.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnot.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnth.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntp.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_compact.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvtnt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_div.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_divr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dot.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eor.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eorv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_expa.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_exth.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_index.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ub.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1ub.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1ub.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsl.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mad.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp32.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_max.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnm.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnmv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_min.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnm.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnmv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mla.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mls.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mmla.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mov.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_msb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mul.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulx.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nand.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_neg.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmad.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmla.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmls.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmsb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nor.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_not.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orn.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orv.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfalse.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfirst.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pnext.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptest.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptrue.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qadd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdech.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecp.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincd.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qinch.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincp.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qsub.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rbit.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rdffr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpe.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recps.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpx.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revb.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revh.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revw.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinta.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinti.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintm.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintn.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintp.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintx.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintz.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrte.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrts.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_scale.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_setffr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sqrt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1b.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1h.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1w.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sub.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_subr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sudot.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tmad.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tsmul.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tssel.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpkhi.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpklo.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_usdot.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilele.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilelt.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_wrffr.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64-bfloat.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64.c
    clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aba.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adalp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlbt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aese.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesimc.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesmc.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bcax.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bdep.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bext.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bgrp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl1n.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl2n.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cdot.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cmla.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtlt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtx.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtxnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eor3.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eorbt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eortb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histcnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histseg.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsub.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsubr.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sh.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sw.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1ub.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uh.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uw.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_logb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_match.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxnmp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minnmp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minp.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mla.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mls.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mul.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nbsl.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nmatch.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmul.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb_128.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt_128.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qabs.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qcadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalbt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslbt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmulh.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qneg.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdcmlah.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlah.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlsh.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmulh.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshl.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshl.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshlu.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsub.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsubr.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rax1.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_recpe.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rhadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshl.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshr.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsqrte.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsra.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sli.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4e.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4ekey.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sqadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sra.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sri.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1b.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1h.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1w.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublbt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subltb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwb.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx-bfloat.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_uqadd.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilege.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilegt.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw-bfloat.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr-bfloat.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr.c
    clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_xar.c

Removed: 
    


################################################################################
diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abd.c
index d1fff7220ccf5..5b5b34d2edca2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,639 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svabd_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabd_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svabd_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svabd_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svabd_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svabd_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabd_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svabd_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svabd_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svabd_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svabd_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabd_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svabd_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabd_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabd_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabd_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabd_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svabd_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svabd_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svabd_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svabd_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabd_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svabd_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabd_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabd_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabd_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabd_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svabd_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svabd_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svabd_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svabd_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svabd_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svabd_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svabd_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svabd_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svabd_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svabd_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svabd_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svabd_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svabd_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svabd_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svabd_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svabd_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svabd_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svabd_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svabd_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svabd_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svabd_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svabd_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svabd_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svabd_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svabd_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svabd_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svabd_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svabd_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svabd_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svabd_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svabd_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svabd_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svabd_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svabd_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svabd_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svabd_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svabd_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svabd_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svabd_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svabd_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svabd_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svabd_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svabd_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svabd_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svabd_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svabd_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svabd_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svabd_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svabd_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svabd_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabd_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svabd_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svabd_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svabd_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svabd_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svabd_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svabd_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svabd_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svabd_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svabd_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svabd_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svabd_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabd_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svabd_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svabd_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabd,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abs.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abs.c
index 26360aca585df..76c5634b91d54 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abs.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_abs.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,188 +14,353 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svabs_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabs_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svabs_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svabs_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabs_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svabs_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabs_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svabs_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabs_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svabs_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabs_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svabs_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svabs_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabs_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svabs_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabs_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svabs_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabs_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svabs_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svabs_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svabs_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svabs_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.abs.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabs_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svabs_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.abs.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabs_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svabs_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.abs.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabs_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svabs_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.abs.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svabs_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svabs_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svabs_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svabs_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svabs_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svabs_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svabs_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svabs_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svabs_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svabs_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svabs_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svabs_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svabs_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svabs_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fabs.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svabs_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svabs_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fabs.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svabs_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabs_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svabs_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svabs_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fabs.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svabs,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acge.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acge.c
index 12ff88c3f4520..791599bd74a37 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acge.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acge.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,54 +14,101 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svacge_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacge_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacge_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svacge_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacge,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacge_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacge_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacge_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svacge_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacge,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacge_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacge_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacge_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svacge_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacge,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacge_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svacge_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svacge_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svacge_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacge,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacge_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svacge_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svacge_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svacge_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacge,_n_f64,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acgt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acgt.c
index 6592ff28e9ac7..b1297fe1cfc27 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acgt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acgt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,54 +14,101 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svacgt_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacgt_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacgt_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svacgt_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facgt.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacgt,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacgt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacgt_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacgt_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svacgt_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacgt,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacgt_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacgt_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacgt_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svacgt_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacgt,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacgt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svacgt_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svacgt_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svacgt_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacgt,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacgt_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svacgt_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svacgt_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svacgt_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacgt,_n_f64,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acle.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acle.c
index f39e73d31c666..b1af88e1c225f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acle.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_acle.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,54 +14,101 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svacle_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacle_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacle_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svacle_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facge.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacle,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacle_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacle_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacle_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svacle_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacle,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacle_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svacle_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svacle_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svacle_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacle,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacle_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svacle_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svacle_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svacle_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[DUP]], <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacle,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svacle_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svacle_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svacle_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svacle_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[DUP]], <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svacle,_n_f64,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_aclt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_aclt.c
index f4b212e5f2aa3..4e5740b498d51 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_aclt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_aclt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,54 +14,101 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaclt_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaclt_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svaclt_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svaclt_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.facgt.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svaclt,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaclt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaclt_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svaclt_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svaclt_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svaclt,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaclt_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaclt_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svaclt_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svaclt_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svaclt,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaclt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaclt_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svaclt_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svaclt_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.facgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[DUP]], <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svaclt,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaclt_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaclt_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svaclt_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svaclt_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.facgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[DUP]], <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svaclt,_n_f64,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_add.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_add.c
index 5c5458686d3df..da95476e2d15c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_add.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_add.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,639 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadd_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadd_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svadd_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svadd_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svadd_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svadd_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadd_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svadd_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svadd_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svadd_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svadd_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadd_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svadd_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svadd_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svadd_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svadd_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadd_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svadd_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svadd_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svadd_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svadd_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadd_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svadd_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svadd_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svadd_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svadd_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadd_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svadd_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svadd_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svadd_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svadd_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svadd_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svadd_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svadd_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svadd_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svadd_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svadd_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svadd_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svadd_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svadd_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svadd_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svadd_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svadd_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svadd_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svadd_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svadd_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svadd_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svadd_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svadd_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svadd_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svadd_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svadd_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svadd_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svadd_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svadd_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svadd_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svadd_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svadd_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svadd_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svadd_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svadd_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svadd_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svadd_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svadd_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svadd_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svadd_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svadd_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svadd_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svadd_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svadd_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svadd_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svadd_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svadd_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svadd_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svadd_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svadd_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svadd_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadd_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svadd_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svadd_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svadd_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svadd_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svadd_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svadd_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svadd_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svadd_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svadd_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svadd_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadd_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadd_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svadd_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svadd_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadd,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adda.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adda.c
index d689ea45d2e7c..fb7ba8e27d978 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adda.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adda.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,29 +14,53 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadda_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fadda.nxv8f16(<vscale x 8 x i1> [[TMP0]], half [[INITIAL:%.*]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadda_f16u10__SVBool_tDhu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fadda.nxv8f16(<vscale x 8 x i1> [[TMP0]], half [[INITIAL:%.*]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svadda_f16(svbool_t pg, float16_t initial, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svadda_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.fadda.nxv8f16(<vscale x 8 x i1> %[[PG]], half %initial, <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadda,_f16,,)(pg, initial, op);
 }
 
+// CHECK-LABEL: @test_svadda_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fadda.nxv4f32(<vscale x 4 x i1> [[TMP0]], float [[INITIAL:%.*]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadda_f32u10__SVBool_tfu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fadda.nxv4f32(<vscale x 4 x i1> [[TMP0]], float [[INITIAL:%.*]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svadda_f32(svbool_t pg, float32_t initial, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svadda_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.fadda.nxv4f32(<vscale x 4 x i1> %[[PG]], float %initial, <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadda,_f32,,)(pg, initial, op);
 }
 
+// CHECK-LABEL: @test_svadda_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fadda.nxv2f64(<vscale x 2 x i1> [[TMP0]], double [[INITIAL:%.*]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svadda_f64u10__SVBool_tdu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fadda.nxv2f64(<vscale x 2 x i1> [[TMP0]], double [[INITIAL:%.*]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svadda_f64(svbool_t pg, float64_t initial, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svadda_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.fadda.nxv2f64(<vscale x 2 x i1> %[[PG]], double %initial, <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadda,_f64,,)(pg, initial, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_addv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_addv.c
index 67b5014832154..21dcd86249187 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_addv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_addv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,99 +14,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddv_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaddv_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svaddv_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svaddv_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.saddv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svaddv_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svaddv_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.saddv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svaddv_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svaddv_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.saddv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.saddv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svaddv_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svaddv_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.saddv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaddv_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svaddv_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svaddv_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svaddv_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svaddv_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svaddv_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svaddv_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svaddv_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svaddv_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uaddv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_u64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.faddv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.faddv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svaddv_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svaddv_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.faddv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.faddv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.faddv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svaddv_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svaddv_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.faddv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svaddv_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.faddv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaddv_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.faddv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svaddv_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svaddv_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.faddv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svaddv,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrb.c
index 12aca6bd18f00..2c8a3ffc77ed8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,34 +14,62 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadrb_u32base_s32offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svadrb_u32base_s32offsetu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrb_u32base_s32offset(svuint32_t bases, svint32_t offsets)
 {
-  // CHECK-LABEL: test_svadrb_u32base_s32offset
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrb_,u32base_s32,offset,)(bases, offsets);
 }
 
+// CHECK-LABEL: @test_svadrb_u64base_s64offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svadrb_u64base_s64offsetu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrb_u64base_s64offset(svuint64_t bases, svint64_t offsets)
 {
-  // CHECK-LABEL: test_svadrb_u64base_s64offset
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64>  %bases, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrb_,u64base_s64,offset,)(bases, offsets);
 }
 
+// CHECK-LABEL: @test_svadrb_u32base_u32offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svadrb_u32base_u32offsetu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrb_u32base_u32offset(svuint32_t bases, svuint32_t offsets)
 {
-  // CHECK-LABEL: test_svadrb_u32base_u32offset
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrb.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrb_,u32base_u32,offset,)(bases, offsets);
 }
 
+// CHECK-LABEL: @test_svadrb_u64base_u64offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svadrb_u64base_u64offsetu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrb_u64base_u64offset(svuint64_t bases, svuint64_t offsets)
 {
-  // CHECK-LABEL: test_svadrb_u64base_u64offset
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrb.nxv2i64(<vscale x 2 x i64>  %bases, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrb_,u64base_u64,offset,)(bases, offsets);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrd.c
index 888d8733704d4..e8438ec7dde84 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,34 +14,62 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadrd_u32base_s32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrd_u32base_s32indexu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrd_u32base_s32index(svuint32_t bases, svint32_t indices)
 {
-  // CHECK-LABEL: test_svadrd_u32base_s32index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrd_,u32base_s32,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrd_u64base_s64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrd_u64base_s64indexu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrd_u64base_s64index(svuint64_t bases, svint64_t indices)
 {
-  // CHECK-LABEL: test_svadrd_u64base_s64index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> %bases, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrd_,u64base_s64,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrd_u32base_u32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrd_u32base_u32indexu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrd_u32base_u32index(svuint32_t bases, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svadrd_u32base_u32index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrd.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrd_,u32base_u32,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrd_u64base_u64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrd_u64base_u64indexu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrd_u64base_u64index(svuint64_t bases, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svadrd_u64base_u64index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrd.nxv2i64(<vscale x 2 x i64> %bases, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrd_,u64base_u64,index,)(bases, indices);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrh.c
index 67e6ebe0f0b28..804e307873b13 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,34 +14,62 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadrh_u32base_s32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrh_u32base_s32indexu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrh_u32base_s32index(svuint32_t bases, svint32_t indices)
 {
-  // CHECK-LABEL: test_svadrh_u32base_s32index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrh_,u32base_s32,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrh_u64base_s64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrh_u64base_s64indexu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrh_u64base_s64index(svuint64_t bases, svint64_t indices)
 {
-  // CHECK-LABEL: test_svadrh_u64base_s64index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> %bases, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrh_,u64base_s64,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrh_u32base_u32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrh_u32base_u32indexu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrh_u32base_u32index(svuint32_t bases, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svadrh_u32base_u32index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrh.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrh_,u32base_u32,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrh_u64base_u64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrh_u64base_u64indexu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrh_u64base_u64index(svuint64_t bases, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svadrh_u64base_u64index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrh.nxv2i64(<vscale x 2 x i64> %bases, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrh_,u64base_u64,index,)(bases, indices);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrw.c
index 226093d6c9b8a..be14505f3751f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_adrw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,34 +14,62 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadrw_u32base_s32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrw_u32base_s32indexu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrw_u32base_s32index(svuint32_t bases, svint32_t indices)
 {
-  // CHECK-LABEL: test_svadrw_u32base_s32index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrw_,u32base_s32,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrw_u64base_s64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrw_u64base_s64indexu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrw_u64base_s64index(svuint64_t bases, svint64_t indices)
 {
-  // CHECK-LABEL: test_svadrw_u64base_s64index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> %bases, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrw_,u64base_s64,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrw_u32base_u32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrw_u32base_u32indexu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> [[BASES:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadrw_u32base_u32index(svuint32_t bases, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svadrw_u32base_u32index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adrw.nxv4i32(<vscale x 4 x i32> %bases, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrw_,u32base_u32,index,)(bases, indices);
 }
 
+// CHECK-LABEL: @test_svadrw_u64base_u64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svadrw_u64base_u64indexu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> [[BASES:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadrw_u64base_u64index(svuint64_t bases, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svadrw_u64base_u64index
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adrw.nxv2i64(<vscale x 2 x i64> %bases, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svadrw_,u64base_u64,index,)(bases, indices);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_and.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_and.c
index 36af1dece3730..33d39cfc2be1f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_and.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_and.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,470 +14,889 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svand_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svand_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svand_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svand_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svand_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svand_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svand_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svand_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svand_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svand_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svand_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svand_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svand_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svand_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svand_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svand_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svand_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svand_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svand_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svand_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svand_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svand_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svand_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svand_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svand_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svand_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svand_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svand_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svand_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svand_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svand_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svand_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svand_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svand_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svand_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svand_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svand_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svand_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svand_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svand_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svand_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svand_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svand_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svand_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svand_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svand_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svand_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svand_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svand_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svand_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svand_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svand_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svand_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svand_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svand_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svand_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svand_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svand_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svand_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svand_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svand_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svand_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svand_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svand_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svand_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svand_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svand_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svand_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svand_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svand_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svand_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svand_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svand_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svand_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svand_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svand_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svand_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svand_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svand_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svand_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svand_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svand_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svand_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svand_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svand_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svand_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svand_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svand_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svand_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svand_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.and.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svand_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.and.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svand_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svand_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.and.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svand,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_andv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_andv.c
index b394b9e7dae4b..0f0a8a7919322 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_andv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_andv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,72 +14,134 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svandv_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.andv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svandv_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.andv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svandv_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svandv_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.andv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.andv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svandv_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.andv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svandv_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svandv_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.andv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.andv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svandv_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.andv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svandv_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svandv_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.andv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.andv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svandv_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.andv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svandv_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svandv_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.andv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.andv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svandv_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.andv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svandv_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svandv_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.andv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.andv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svandv_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.andv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svandv_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svandv_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.andv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.andv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svandv_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.andv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svandv_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svandv_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.andv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svandv_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.andv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svandv_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.andv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svandv_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svandv_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.andv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svandv,_u64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asr.c
index 20b12b6f6fabf..b4bb06a44c39f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asr.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,403 +14,762 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svasr_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svasr_s8_zu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasr_s8_z(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svasr_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s16_zu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasr_s16_z(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svasr_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s32_zu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasr_s32_z(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svasr_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s64_zu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svasr_s64_z(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svasr_s8_mu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svasr_s8_m(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svasr_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s16_mu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svasr_s16_m(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svasr_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s32_mu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svasr_s32_m(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svasr_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s64_mu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svasr_s64_m(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svasr_s8_xu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svasr_s8_x(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svasr_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s16_xu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svasr_s16_x(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svasr_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s32_xu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svasr_s32_x(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svasr_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svasr_s64_xu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svasr_s64_x(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s64_zu10__SVBool_tu11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svasr_n_s64_z(svbool_t pg, svint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s64_mu10__SVBool_tu11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svasr_n_s64_m(svbool_t pg, svint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s64_xu10__SVBool_tu11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svasr_n_s64_x(svbool_t pg, svint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svasr_wide_s8_zu10__SVBool_tu10__SVInt8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasr_wide_s8_z(svbool_t pg, svint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasr_wide_s16_zu10__SVBool_tu11__SVInt16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasr_wide_s16_z(svbool_t pg, svint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasr_wide_s32_zu10__SVBool_tu11__SVInt32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasr_wide_s32_z(svbool_t pg, svint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svasr_wide_s8_mu10__SVBool_tu10__SVInt8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svasr_wide_s8_m(svbool_t pg, svint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasr_wide_s16_mu10__SVBool_tu11__SVInt16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svasr_wide_s16_m(svbool_t pg, svint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasr_wide_s32_mu10__SVBool_tu11__SVInt32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svasr_wide_s32_m(svbool_t pg, svint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svasr_wide_s8_xu10__SVBool_tu10__SVInt8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svasr_wide_s8_x(svbool_t pg, svint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasr_wide_s16_xu10__SVBool_tu11__SVInt16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svasr_wide_s16_x(svbool_t pg, svint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasr_wide_s32_xu10__SVBool_tu11__SVInt32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svasr_wide_s32_x(svbool_t pg, svint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svasr_n_s8_zu10__SVBool_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svasr_n_s8_z(svbool_t pg, svint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s16_zu10__SVBool_tu11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svasr_n_s16_z(svbool_t pg, svint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s32_zu10__SVBool_tu11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svasr_n_s32_z(svbool_t pg, svint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svasr_n_s8_mu10__SVBool_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasr_n_s8_m(svbool_t pg, svint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s16_mu10__SVBool_tu11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasr_n_s16_m(svbool_t pg, svint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s32_mu10__SVBool_tu11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasr_n_s32_m(svbool_t pg, svint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svasr_n_s8_xu10__SVBool_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasr_n_s8_x(svbool_t pg, svint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s16_xu10__SVBool_tu11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasr_n_s16_x(svbool_t pg, svint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasr_n_s32_xu10__SVBool_tu11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasr_n_s32_x(svbool_t pg, svint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svasr_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svasr_wide_n_s8_mu10__SVBool_tu10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasr_wide_n_s8_m(svbool_t pg, svint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svasr_wide_n_s16_mu10__SVBool_tu11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasr_wide_n_s16_m(svbool_t pg, svint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svasr_wide_n_s32_mu10__SVBool_tu11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasr_wide_n_s32_m(svbool_t pg, svint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svasr_wide_n_s8_zu10__SVBool_tu10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svasr_wide_n_s8_z(svbool_t pg, svint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[PG]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svasr_wide_n_s16_zu10__SVBool_tu11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svasr_wide_n_s16_z(svbool_t pg, svint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[OP]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svasr_wide_n_s32_zu10__SVBool_tu11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svasr_wide_n_s32_z(svbool_t pg, svint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[OP]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svasr_wide_n_s8_xu10__SVBool_tu10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasr_wide_n_s8_x(svbool_t pg, svint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svasr_wide_n_s16_xu10__SVBool_tu11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasr_wide_n_s16_x(svbool_t pg, svint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svasr_wide_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svasr_wide_n_s32_xu10__SVBool_tu11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasr_wide_n_s32_x(svbool_t pg, svint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svasr_wide_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasr_wide,_n_s32,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asrd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asrd.c
index 2de1b048c5134..7efe8ac98717f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asrd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_asrd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,150 +14,282 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svasrd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasrd_n_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasrd_n_s8_z(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s8,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s8_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svasrd_n_s8_z_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svasrd_n_s8_z_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s8_z_1
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s8,_z,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasrd_n_s16_z(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s16,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s16_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasrd_n_s16_z_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svasrd_n_s16_z_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s16_z_1
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s16,_z,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasrd_n_s32_z(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s32,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s32_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasrd_n_s32_z_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svasrd_n_s32_z_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s32_z_1
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s32,_z,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svasrd_n_s64_z(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s64,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s64_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svasrd_n_s64_z_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svasrd_n_s64_z_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s64_z_1
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s64,_z,)(pg, op1, 64);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasrd_n_s8_mu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svasrd_n_s8_m(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s8,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s16_mu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svasrd_n_s16_m(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s16,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s32_mu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svasrd_n_s32_m(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s32,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s64_mu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svasrd_n_s64_m(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s64,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svasrd_n_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svasrd_n_s8_x(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.asrd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s8,_x,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svasrd_n_s16_x(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.asrd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s16,_x,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svasrd_n_s32_x(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.asrd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s32,_x,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svasrd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svasrd_n_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svasrd_n_s64_x(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svasrd_n_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.asrd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svasrd,_n_s64,_x,)(pg, op1, 64);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfdot.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfdot.c
index 62258237bc5ef..deab191390c00 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfdot.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfdot.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,31 +13,60 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_bfdot_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_bfdot_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfdot_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfdot_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfdot, _f32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_bfdot_lane_0_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_bfdot_lane_0_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfdot_lane_0_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfdot_lane_0_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot.lane(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z, i64 0)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfdot_lane, _f32, , )(x, y, z, 0);
 }
 
+// CHECK-LABEL: @test_bfdot_lane_3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 3)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_bfdot_lane_3_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfdot_lane_3_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfdot_lane_3_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot.lane(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z, i64 3)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfdot_lane, _f32, , )(x, y, z, 3);
 }
 
+// CHECK-LABEL: @test_bfdot_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[Z:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_bfdot_n_f32u13__SVFloat32_tu14__SVBFloat16_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[Z:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_bfdot_n_f32(svfloat32_t x, svbfloat16_t y, bfloat16_t z) {
-  // CHECK-LABEL: test_bfdot_n_f32
-  // CHECK: %[[SPLAT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat %z)
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfdot(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %[[SPLAT]])
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfdot, _n_f32, , )(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalb.c
index 8669de8ea3224..c5733eb812e7f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16  -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16  -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16  -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,31 +13,60 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbfmlalb_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbfmlalb_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svbfmlalb_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_svbfmlalb_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalb, _f32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_bfmlalb_lane_0_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_bfmlalb_lane_0_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfmlalb_lane_0_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfmlalb_lane_0_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb.lane(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z, i64 0)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalb_lane, _f32, , )(x, y, z, 0);
 }
 
+// CHECK-LABEL: @test_bfmlalb_lane_7_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 7)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_bfmlalb_lane_7_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfmlalb_lane_7_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfmlalb_lane_7_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb.lane(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z, i64 7)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalb_lane, _f32, , )(x, y, z, 7);
 }
 
+// CHECK-LABEL: @test_bfmlalb_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[Z:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_bfmlalb_n_f32u13__SVFloat32_tu14__SVBFloat16_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[Z:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_bfmlalb_n_f32(svfloat32_t x, svbfloat16_t y, bfloat16_t z) {
-  // CHECK-LABEL: test_bfmlalb_n_f32
-  // CHECK: %[[SPLAT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat %z)
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalb(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %[[SPLAT]])
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalb, _n_f32, , )(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalt.c
index 006ff6bc3200b..6fdd3206fee6b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmlalt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,31 +13,60 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbfmlalt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbfmlalt_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svbfmlalt_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_svbfmlalt_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalt, _f32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_bfmlalt_lane_0_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_bfmlalt_lane_0_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfmlalt_lane_0_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfmlalt_lane_0_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt.lane(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z, i64 0)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalt_lane, _f32, , )(x, y, z, 0);
 }
 
+// CHECK-LABEL: @test_bfmlalt_lane_7_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 7)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_bfmlalt_lane_7_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt.lane(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]], i64 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfmlalt_lane_7_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfmlalt_lane_7_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt.lane(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z, i64 7)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalt_lane, _f32, , )(x, y, z, 7);
 }
 
+// CHECK-LABEL: @test_bfmlalt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[Z:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_bfmlalt_n_f32u13__SVFloat32_tu14__SVBFloat16_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[Z:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_bfmlalt_n_f32(svfloat32_t x, svbfloat16_t y, bfloat16_t z) {
-  // CHECK-LABEL: test_bfmlalt_n_f32
-  // CHECK: %[[SPLAT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat %z)
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmlalt(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %[[SPLAT]])
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmlalt, _n_f32, , )(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmmla.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmmla.c
index 0d5af49d27582..c7f059f7904ab 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmmla.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bfmmla.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,9 +13,16 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_bfmmla_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmmla(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_bfmmla_f32u13__SVFloat32_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmmla(<vscale x 4 x float> [[X:%.*]], <vscale x 8 x bfloat> [[Y:%.*]], <vscale x 8 x bfloat> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_bfmmla_f32(svfloat32_t x, svbfloat16_t y, svbfloat16_t z) {
-  // CHECK-LABEL: test_bfmmla_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.bfmmla(<vscale x 4 x float> %x, <vscale x 8 x bfloat> %y, <vscale x 8 x bfloat> %z)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svbfmmla, _f32, , )(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bic.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bic.c
index fb11e485ca87a..e391d41fe795b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bic.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_bic.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,470 +14,889 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbic_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbic_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbic_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svbic_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svbic_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svbic_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbic_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbic_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svbic_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svbic_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svbic_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbic_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svbic_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svbic_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svbic_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svbic_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbic_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbic_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbic_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbic_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbic_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbic_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svbic_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svbic_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svbic_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svbic_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbic_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbic_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbic_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbic_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbic_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbic_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbic_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svbic_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svbic_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svbic_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svbic_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbic_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svbic_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svbic_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svbic_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svbic_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbic_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbic_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svbic_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svbic_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svbic_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbic_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbic_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svbic_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svbic_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svbic_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbic_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbic_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svbic_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svbic_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svbic_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbic_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbic_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svbic_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svbic_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbic_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svbic_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svbic_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svbic_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.bic.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbic_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.bic.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbic_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svbic_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.bic.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbic,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brka.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brka.c
index 2e3b1d798d6da..3bccc717c5f01 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brka.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brka.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,18 +14,32 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbrka_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brka.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbrka_b_zu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brka.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrka_b_z(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svbrka_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brka.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrka,_b,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svbrka_b_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brka.nxv16i1(<vscale x 16 x i1> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbrka_b_mu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brka.nxv16i1(<vscale x 16 x i1> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrka_b_m(svbool_t inactive, svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svbrka_b_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brka.nxv16i1(<vscale x 16 x i1> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrka,_b,_m,)(inactive, pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkb.c
index 239f7898f8638..6e4da6dad7af5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,18 +14,32 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbrkb_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkb.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbrkb_b_zu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkb.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrkb_b_z(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svbrkb_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkb.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrkb,_b,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svbrkb_b_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkb.nxv16i1(<vscale x 16 x i1> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbrkb_b_mu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkb.nxv16i1(<vscale x 16 x i1> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrkb_b_m(svbool_t inactive, svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svbrkb_b_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkb.nxv16i1(<vscale x 16 x i1> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrkb,_b,_m,)(inactive, pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkn.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkn.c
index c7e5d3fdccf7a..1c9d7910eacbe 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkn.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkn.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbrkn_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkn.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbrkn_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkn.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrkn_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svbrkn_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkn.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrkn,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpa.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpa.c
index b3c0d3a05b5da..9597d0fffe1cb 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpa.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpa.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbrkpa_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkpa.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbrkpa_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkpa.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrkpa_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svbrkpa_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkpa.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrkpa,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpb.c
index 8bd9500bffb90..2fe10fd9d805e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_brkpb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbrkpb_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkpb.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbrkpb_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkpb.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svbrkpb_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svbrkpb_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.brkpb.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svbrkpb,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cadd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cadd.c
index ff61eef172c2f..21500914e9f13 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cadd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cadd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,96 +14,180 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcadd_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svcadd_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, i32 90)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f16,_z,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f16_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcadd_f16_z_1u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svcadd_f16_z_1(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f16_z_1
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, i32 270)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f16,_z,)(pg, op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svcadd_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, i32 90)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f32,_z,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svcadd_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, i32 90)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f64,_z,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcadd_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 90)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f16,_m,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcadd_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, i32 90)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f32,_m,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcadd_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, i32 90)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f64,_m,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcadd_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcadd.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 90)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f16,_x,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcadd_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcadd.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, i32 90)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f32,_x,)(pg, op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcadd_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcadd.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, i32 90)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcadd,_f64,_x,)(pg, op1, op2, 90);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta-bfloat.c
index 21e11d294e8f8..98e7cea30bcb0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,20 +15,36 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svclasta_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.clasta.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svclasta_bf16u10__SVBool_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.clasta.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svclasta_bf16(svbool_t pg, svbfloat16_t fallback, svbfloat16_t data) {
-  // CHECK-LABEL: test_svclasta_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.clasta.nxv8bf16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %fallback, <vscale x 8 x bfloat> %data)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svclasta_bf16'}}
   return SVE_ACLE_FUNC(svclasta, _bf16, , )(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.clasta.n.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CHECK-NEXT:    ret bfloat [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svclasta_n_bf16u10__SVBool_tu6__bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.clasta.n.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret bfloat [[TMP1]]
+//
 bfloat16_t test_svclasta_n_bf16(svbool_t pg, bfloat16_t fallback, svbfloat16_t data) {
-  // CHECK-LABEL: test_svclasta_n_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call bfloat @llvm.aarch64.sve.clasta.n.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat %fallback, <vscale x 8 x bfloat> %data)
-  // CHECK: ret bfloat %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svclasta_n_bf16'}}
   return SVE_ACLE_FUNC(svclasta, _n_bf16, , )(pg, fallback, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta.c
index 009af721dd3cb..c6cf55accf820 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clasta.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,196 +14,368 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svclasta_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclasta_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svclasta_s8(svbool_t pg, svint8_t fallback, svint8_t data)
 {
-  // CHECK-LABEL: test_svclasta_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clasta.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_s8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svclasta_s16(svbool_t pg, svint16_t fallback, svint16_t data)
 {
-  // CHECK-LABEL: test_svclasta_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clasta.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_s16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svclasta_s32(svbool_t pg, svint32_t fallback, svint32_t data)
 {
-  // CHECK-LABEL: test_svclasta_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clasta.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_s32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svclasta_s64(svbool_t pg, svint64_t fallback, svint64_t data)
 {
-  // CHECK-LABEL: test_svclasta_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clasta.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_s64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclasta_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclasta_u8(svbool_t pg, svuint8_t fallback, svuint8_t data)
 {
-  // CHECK-LABEL: test_svclasta_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clasta.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_u8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclasta_u16(svbool_t pg, svuint16_t fallback, svuint16_t data)
 {
-  // CHECK-LABEL: test_svclasta_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clasta.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_u16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclasta_u32(svbool_t pg, svuint32_t fallback, svuint32_t data)
 {
-  // CHECK-LABEL: test_svclasta_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clasta.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_u32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclasta_u64(svbool_t pg, svuint64_t fallback, svuint64_t data)
 {
-  // CHECK-LABEL: test_svclasta_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clasta.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_u64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.clasta.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.clasta.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svclasta_f16(svbool_t pg, svfloat16_t fallback, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svclasta_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.clasta.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %fallback, <vscale x 8 x half> %data)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_f16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.clasta.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.clasta.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svclasta_f32(svbool_t pg, svfloat32_t fallback, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svclasta_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.clasta.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %fallback, <vscale x 4 x float> %data)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_f32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.clasta.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclasta_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.clasta.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svclasta_f64(svbool_t pg, svfloat64_t fallback, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svclasta_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.clasta.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %fallback, <vscale x 2 x double> %data)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_f64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clasta.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svclasta_n_s8u10__SVBool_tau10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clasta.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svclasta_n_s8(svbool_t pg, int8_t fallback, svint8_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.clasta.n.nxv16i8(<vscale x 16 x i1> %pg, i8 %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_s8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clasta.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_s16u10__SVBool_tsu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clasta.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svclasta_n_s16(svbool_t pg, int16_t fallback, svint16_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.clasta.n.nxv8i16(<vscale x 8 x i1> %[[PG]], i16 %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_s16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clasta.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_s32u10__SVBool_tiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clasta.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svclasta_n_s32(svbool_t pg, int32_t fallback, svint32_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.clasta.n.nxv4i32(<vscale x 4 x i1> %[[PG]], i32 %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_s32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clasta.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_s64u10__SVBool_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clasta.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svclasta_n_s64(svbool_t pg, int64_t fallback, svint64_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.clasta.n.nxv2i64(<vscale x 2 x i1> %[[PG]], i64 %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_s64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clasta.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svclasta_n_u8u10__SVBool_thu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clasta.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svclasta_n_u8(svbool_t pg, uint8_t fallback, svuint8_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.clasta.n.nxv16i8(<vscale x 16 x i1> %pg, i8 %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_u8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clasta.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_u16u10__SVBool_ttu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clasta.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svclasta_n_u16(svbool_t pg, uint16_t fallback, svuint16_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.clasta.n.nxv8i16(<vscale x 8 x i1> %[[PG]], i16 %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_u16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clasta.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_u32u10__SVBool_tju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clasta.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svclasta_n_u32(svbool_t pg, uint32_t fallback, svuint32_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.clasta.n.nxv4i32(<vscale x 4 x i1> %[[PG]], i32 %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_u32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clasta.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_u64u10__SVBool_tmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clasta.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svclasta_n_u64(svbool_t pg, uint64_t fallback, svuint64_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.clasta.n.nxv2i64(<vscale x 2 x i1> %[[PG]], i64 %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_u64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.clasta.n.nxv8f16(<vscale x 8 x i1> [[TMP0]], half [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_f16u10__SVBool_tDhu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.clasta.n.nxv8f16(<vscale x 8 x i1> [[TMP0]], half [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svclasta_n_f16(svbool_t pg, float16_t fallback, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.clasta.n.nxv8f16(<vscale x 8 x i1> %[[PG]], half %fallback, <vscale x 8 x half> %data)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_f16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.clasta.n.nxv4f32(<vscale x 4 x i1> [[TMP0]], float [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_f32u10__SVBool_tfu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.clasta.n.nxv4f32(<vscale x 4 x i1> [[TMP0]], float [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svclasta_n_f32(svbool_t pg, float32_t fallback, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.clasta.n.nxv4f32(<vscale x 4 x i1> %[[PG]], float %fallback, <vscale x 4 x float> %data)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_f32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclasta_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.clasta.n.nxv2f64(<vscale x 2 x i1> [[TMP0]], double [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclasta_n_f64u10__SVBool_tdu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.clasta.n.nxv2f64(<vscale x 2 x i1> [[TMP0]], double [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svclasta_n_f64(svbool_t pg, float64_t fallback, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svclasta_n_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.clasta.n.nxv2f64(<vscale x 2 x i1> %[[PG]], double %fallback, <vscale x 2 x double> %data)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclasta,_n_f64,,)(pg, fallback, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb-bfloat.c
index cf16ddc134fb8..0bd0c3b859860 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,20 +15,36 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svclastb_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.clastb.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svclastb_bf16u10__SVBool_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.clastb.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svclastb_bf16(svbool_t pg, svbfloat16_t fallback, svbfloat16_t data) {
-  // CHECK-LABEL: test_svclastb_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.clastb.nxv8bf16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %fallback, <vscale x 8 x bfloat> %data)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svclastb_bf16'}}
   return SVE_ACLE_FUNC(svclastb, _bf16, , )(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.clastb.n.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CHECK-NEXT:    ret bfloat [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svclastb_n_bf16u10__SVBool_tu6__bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.clastb.n.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret bfloat [[TMP1]]
+//
 bfloat16_t test_svclastb_n_bf16(svbool_t pg, bfloat16_t fallback, svbfloat16_t data) {
-  // CHECK-LABEL: test_svclastb_n_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call bfloat @llvm.aarch64.sve.clastb.n.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat %fallback, <vscale x 8 x bfloat> %data)
-  // CHECK: ret bfloat %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svclastb_n_bf16'}}
   return SVE_ACLE_FUNC(svclastb, _n_bf16, , )(pg, fallback, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb.c
index 64c393ffb2d01..1473d0ab62070 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clastb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,196 +14,368 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svclastb_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclastb_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svclastb_s8(svbool_t pg, svint8_t fallback, svint8_t data)
 {
-  // CHECK-LABEL: test_svclastb_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clastb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_s8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svclastb_s16(svbool_t pg, svint16_t fallback, svint16_t data)
 {
-  // CHECK-LABEL: test_svclastb_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clastb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_s16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svclastb_s32(svbool_t pg, svint32_t fallback, svint32_t data)
 {
-  // CHECK-LABEL: test_svclastb_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clastb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_s32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svclastb_s64(svbool_t pg, svint64_t fallback, svint64_t data)
 {
-  // CHECK-LABEL: test_svclastb_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clastb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_s64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclastb_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclastb_u8(svbool_t pg, svuint8_t fallback, svuint8_t data)
 {
-  // CHECK-LABEL: test_svclastb_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clastb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_u8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclastb_u16(svbool_t pg, svuint16_t fallback, svuint16_t data)
 {
-  // CHECK-LABEL: test_svclastb_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clastb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_u16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclastb_u32(svbool_t pg, svuint32_t fallback, svuint32_t data)
 {
-  // CHECK-LABEL: test_svclastb_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clastb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_u32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclastb_u64(svbool_t pg, svuint64_t fallback, svuint64_t data)
 {
-  // CHECK-LABEL: test_svclastb_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clastb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_u64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.clastb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.clastb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svclastb_f16(svbool_t pg, svfloat16_t fallback, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svclastb_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.clastb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %fallback, <vscale x 8 x half> %data)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_f16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.clastb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.clastb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svclastb_f32(svbool_t pg, svfloat32_t fallback, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svclastb_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.clastb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %fallback, <vscale x 4 x float> %data)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_f32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.clastb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svclastb_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.clastb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svclastb_f64(svbool_t pg, svfloat64_t fallback, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svclastb_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.clastb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %fallback, <vscale x 2 x double> %data)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_f64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clastb.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svclastb_n_s8u10__SVBool_tau10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clastb.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svclastb_n_s8(svbool_t pg, int8_t fallback, svint8_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.clastb.n.nxv16i8(<vscale x 16 x i1> %pg, i8 %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_s8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clastb.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_s16u10__SVBool_tsu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clastb.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svclastb_n_s16(svbool_t pg, int16_t fallback, svint16_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.clastb.n.nxv8i16(<vscale x 8 x i1> %[[PG]], i16 %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_s16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clastb.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_s32u10__SVBool_tiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clastb.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svclastb_n_s32(svbool_t pg, int32_t fallback, svint32_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.clastb.n.nxv4i32(<vscale x 4 x i1> %[[PG]], i32 %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_s32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clastb.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_s64u10__SVBool_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clastb.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svclastb_n_s64(svbool_t pg, int64_t fallback, svint64_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.clastb.n.nxv2i64(<vscale x 2 x i1> %[[PG]], i64 %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_s64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clastb.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svclastb_n_u8u10__SVBool_thu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.clastb.n.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8 [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svclastb_n_u8(svbool_t pg, uint8_t fallback, svuint8_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.clastb.n.nxv16i8(<vscale x 16 x i1> %pg, i8 %fallback, <vscale x 16 x i8> %data)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_u8,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clastb.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_u16u10__SVBool_ttu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.clastb.n.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16 [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svclastb_n_u16(svbool_t pg, uint16_t fallback, svuint16_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.clastb.n.nxv8i16(<vscale x 8 x i1> %[[PG]], i16 %fallback, <vscale x 8 x i16> %data)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_u16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clastb.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_u32u10__SVBool_tju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.clastb.n.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32 [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svclastb_n_u32(svbool_t pg, uint32_t fallback, svuint32_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.clastb.n.nxv4i32(<vscale x 4 x i1> %[[PG]], i32 %fallback, <vscale x 4 x i32> %data)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_u32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clastb.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_u64u10__SVBool_tmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.clastb.n.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64 [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svclastb_n_u64(svbool_t pg, uint64_t fallback, svuint64_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.clastb.n.nxv2i64(<vscale x 2 x i1> %[[PG]], i64 %fallback, <vscale x 2 x i64> %data)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_u64,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.clastb.n.nxv8f16(<vscale x 8 x i1> [[TMP0]], half [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_f16u10__SVBool_tDhu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.clastb.n.nxv8f16(<vscale x 8 x i1> [[TMP0]], half [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svclastb_n_f16(svbool_t pg, float16_t fallback, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.clastb.n.nxv8f16(<vscale x 8 x i1> %[[PG]], half %fallback, <vscale x 8 x half> %data)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_f16,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.clastb.n.nxv4f32(<vscale x 4 x i1> [[TMP0]], float [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_f32u10__SVBool_tfu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.clastb.n.nxv4f32(<vscale x 4 x i1> [[TMP0]], float [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svclastb_n_f32(svbool_t pg, float32_t fallback, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.clastb.n.nxv4f32(<vscale x 4 x i1> %[[PG]], float %fallback, <vscale x 4 x float> %data)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_f32,,)(pg, fallback, data);
 }
 
+// CHECK-LABEL: @test_svclastb_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.clastb.n.nxv2f64(<vscale x 2 x i1> [[TMP0]], double [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svclastb_n_f64u10__SVBool_tdu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.clastb.n.nxv2f64(<vscale x 2 x i1> [[TMP0]], double [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svclastb_n_f64(svbool_t pg, float64_t fallback, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svclastb_n_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.clastb.n.nxv2f64(<vscale x 2 x i1> %[[PG]], double %fallback, <vscale x 2 x double> %data)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclastb,_n_f64,,)(pg, fallback, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cls.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cls.c
index c1f159dc73656..c47252f7bd7e3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cls.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cls.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,107 +14,200 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcls_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcls_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcls_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcls_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcls_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcls_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcls_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcls_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcls_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcls_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcls_s8_mu11__SVUint8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcls_s8_m(svuint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcls_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s16_mu12__SVUint16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcls_s16_m(svuint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcls_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s32_mu12__SVUint32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcls_s32_m(svuint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcls_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s64_mu12__SVUint64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcls_s64_m(svuint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcls_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcls_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcls_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcls_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cls.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcls_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcls_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cls.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcls_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcls_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cls.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcls_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcls_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcls_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcls_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cls.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcls,_s64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clz.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clz.c
index ee64519b9fe0b..51e29598a9b55 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clz.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_clz.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,212 +14,398 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svclz_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svclz_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclz_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svclz_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclz_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svclz_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclz_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svclz_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclz_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svclz_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svclz_u8_zu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclz_u8_z(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svclz_u8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclz_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svclz_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclz_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svclz_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclz_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svclz_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svclz_s8_mu11__SVUint8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclz_s8_m(svuint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svclz_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s16_mu12__SVUint16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclz_s16_m(svuint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svclz_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s32_mu12__SVUint32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclz_s32_m(svuint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svclz_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s64_mu12__SVUint64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclz_s64_m(svuint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svclz_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svclz_u8_mu11__SVUint8_tu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclz_u8_m(svuint8_t inactive, svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svclz_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclz_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svclz_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclz_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svclz_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclz_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svclz_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svclz_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclz_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svclz_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclz_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svclz_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclz_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svclz_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclz_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svclz_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svclz_u8_xu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svclz_u8_x(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svclz_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.clz.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svclz_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svclz_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.clz.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svclz_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svclz_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.clz.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svclz_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svclz_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svclz_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svclz_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.clz.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svclz,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmla.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmla.c
index 0c36e69853ed2..0ffa0cb09de25 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmla.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmla.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,148 +14,278 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmla_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svcmla_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f16,_z,)(pg, op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_f16_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcmla_f16_z_1u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svcmla_f16_z_1(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f16_z_1
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 90)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f16,_z,)(pg, op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_f16_z_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcmla_f16_z_2u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svcmla_f16_z_2(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f16_z_2
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 180)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f16,_z,)(pg, op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_f16_z_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcmla_f16_z_3u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svcmla_f16_z_3(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f16_z_3
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 270)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f16,_z,)(pg, op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svcmla_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f32,_z,)(pg, op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svcmla_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 90)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f64,_z,)(pg, op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcmla_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 180)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f16,_m,)(pg, op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcmla_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 270)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f32,_m,)(pg, op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcmla_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f64,_m,)(pg, op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcmla_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 90)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f16,_x,)(pg, op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcmla_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 180)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f32,_x,)(pg, op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcmla_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 270)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla,_f64,_x,)(pg, op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0, i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmla_lane_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0, i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svcmla_lane_f16(svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0, i32 0)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla_lane,_f16,,)(op1, op2, op3, 0, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_f16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 3, i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmla_lane_f16_1u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 3, i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svcmla_lane_f16_1(svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_f16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcmla.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 3, i32 90)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla_lane,_f16,,)(op1, op2, op3, 3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0, i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmla_lane_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0, i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svcmla_lane_f32(svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 0, i32 180)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla_lane,_f32,,)(op1, op2, op3, 0, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 1, i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmla_lane_f32_1u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 1, i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svcmla_lane_f32_1(svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcmla.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 1, i32 270)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmla_lane,_f32,,)(op1, op2, op3, 1, 270);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpeq.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpeq.c
index 3c267b3c49c3b..cedd2738b7ef7 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpeq.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpeq.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,284 +14,538 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmpeq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpeq_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpeq_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpeq,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpeq_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpeq_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpeq,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_s64u10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_s64(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_u64u10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_u64(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpeq.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_wide_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmpeq_wide_s8u10__SVBool_tu10__SVInt8_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpeq_wide_s8(svbool_t pg, svint8_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_wide_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpeq_wide,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_wide_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpeq_wide_s16u10__SVBool_tu11__SVInt16_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_wide_s16(svbool_t pg, svint16_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_wide_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq_wide,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_wide_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpeq_wide_s32u10__SVBool_tu11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_wide_s32(svbool_t pg, svint32_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_wide_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq_wide,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpeq_n_s8u10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpeq_n_s8(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpeq,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_s16u10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_s16(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_s32u10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_s32(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpeq_n_u8u10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpeq_n_u8(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpeq,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_u16u10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_u16(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_u32u10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_u32(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpeq.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpeq.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpeq.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpeq.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpeq.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpeq.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpeq.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpeq_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpeq.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpeq_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpeq.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpeq.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpeq.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpeq.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpeq.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpeq.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpeq.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpeq.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpeq_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpeq.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpeq.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq,_n_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_wide_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmpeq_wide_n_s8u10__SVBool_tu10__SVInt8_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpeq_wide_n_s8(svbool_t pg, svint8_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_wide_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpeq_wide,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_wide_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpeq_wide_n_s16u10__SVBool_tu11__SVInt16_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_wide_n_s16(svbool_t pg, svint16_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_wide_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq_wide,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpeq_wide_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpeq_wide_n_s32u10__SVBool_tu11__SVInt32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpeq_wide_n_s32(svbool_t pg, svint32_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpeq_wide_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpeq.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpeq_wide,_n_s32,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpge.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpge.c
index 895e49c69af22..26ead4897ea17 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpge.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpge.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,343 +14,650 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmpge_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpge_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpge_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpge_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpge_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_s64u10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_s64(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_u64u10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_u64(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmpge_wide_s8u10__SVBool_tu10__SVInt8_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpge_wide_s8(svbool_t pg, svint8_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge_wide,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpge_wide_s16u10__SVBool_tu11__SVInt16_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_wide_s16(svbool_t pg, svint16_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpge_wide_s32u10__SVBool_tu11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_wide_s32(svbool_t pg, svint32_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmpge_wide_u8u10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpge_wide_u8(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge_wide,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpge_wide_u16u10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_wide_u16(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpge_wide_u32u10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_wide_u32(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpge_n_s8u10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpge_n_s8(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_s16u10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_s16(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_s32u10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_s32(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpge_n_u8u10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpge_n_u8(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_u16u10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_u16(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_u32u10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_u32(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpge_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpge_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpge_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge,_n_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmpge_wide_n_s8u10__SVBool_tu10__SVInt8_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpge_wide_n_s8(svbool_t pg, svint8_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge_wide,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpge_wide_n_s16u10__SVBool_tu11__SVInt16_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_wide_n_s16(svbool_t pg, svint16_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpge_wide_n_s32u10__SVBool_tu11__SVInt32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_wide_n_s32(svbool_t pg, svint32_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmpge_wide_n_u8u10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpge_wide_n_u8(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpge_wide,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpge_wide_n_u16u10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_wide_n_u16(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpge_wide_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpge_wide_n_u32u10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpge_wide_n_u32(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpge_wide_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpge_wide,_n_u32,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpgt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpgt.c
index 9d11062448b26..a5881cb129ab1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpgt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpgt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,343 +14,650 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmpgt_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpgt_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpgt_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpgt_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpgt_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_s64u10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_s64(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_u64u10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_u64(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmpgt_wide_s8u10__SVBool_tu10__SVInt8_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpgt_wide_s8(svbool_t pg, svint8_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpgt_wide_s16u10__SVBool_tu11__SVInt16_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_wide_s16(svbool_t pg, svint16_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpgt_wide_s32u10__SVBool_tu11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_wide_s32(svbool_t pg, svint32_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmpgt_wide_u8u10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpgt_wide_u8(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpgt_wide_u16u10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_wide_u16(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpgt_wide_u32u10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_wide_u32(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpgt_n_s8u10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpgt_n_s8(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_s16u10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_s16(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_s32u10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_s32(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpgt_n_u8u10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpgt_n_u8(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_u16u10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_u16(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_u32u10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_u32(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpgt_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpgt_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpgt_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt,_n_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmpgt_wide_n_s8u10__SVBool_tu10__SVInt8_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpgt_wide_n_s8(svbool_t pg, svint8_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpgt_wide_n_s16u10__SVBool_tu11__SVInt16_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_wide_n_s16(svbool_t pg, svint16_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpgt_wide_n_s32u10__SVBool_tu11__SVInt32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_wide_n_s32(svbool_t pg, svint32_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmpgt_wide_n_u8u10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpgt_wide_n_u8(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpgt_wide_n_u16u10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_wide_n_u16(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpgt_wide_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpgt_wide_n_u32u10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpgt_wide_n_u32(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpgt_wide_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpgt_wide,_n_u32,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmple.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmple.c
index c3f6408b1126f..694b7c7a93afc 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmple.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmple.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,343 +14,650 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmple_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmple_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmple_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcmple_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcmple_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcmple_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmple_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmple_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcmple_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcmple_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcmple_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_s64u10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_s64(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpge.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[DUP]], <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_u64u10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_u64(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphs.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[DUP]], <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmple.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmple_wide_s8u10__SVBool_tu10__SVInt8_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmple.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmple_wide_s8(svbool_t pg, svint8_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmple.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple_wide,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmple.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmple_wide_s16u10__SVBool_tu11__SVInt16_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmple.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_wide_s16(svbool_t pg, svint16_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmple.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmple.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmple_wide_s32u10__SVBool_tu11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmple.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_wide_s32(svbool_t pg, svint32_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmple.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpls.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmple_wide_u8u10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpls.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmple_wide_u8(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpls.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple_wide,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpls.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmple_wide_u16u10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpls.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_wide_u16(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpls.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpls.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmple_wide_u32u10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpls.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_wide_u32(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpls.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmple_n_s8u10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmple_n_s8(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpge.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[DUP]], <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_s16u10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_s16(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpge.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[DUP]], <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_s32u10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_s32(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpge.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[DUP]], <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmple_n_u8u10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmple_n_u8(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphs.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[DUP]], <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_u16u10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_u16(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphs.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[DUP]], <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_u32u10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_u32(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphs.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[DUP]], <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmple_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmple_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmple_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmple_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpge.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[DUP]], <vscale x 8 x half> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpge.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[DUP]], <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmple_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpge.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[DUP]], <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple,_n_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmple.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmple_wide_n_s8u10__SVBool_tu10__SVInt8_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmple.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmple_wide_n_s8(svbool_t pg, svint8_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmple.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple_wide,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmple.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmple_wide_n_s16u10__SVBool_tu11__SVInt16_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmple.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_wide_n_s16(svbool_t pg, svint16_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmple.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmple.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmple_wide_n_s32u10__SVBool_tu11__SVInt32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmple.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_wide_n_s32(svbool_t pg, svint32_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmple.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpls.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmple_wide_n_u8u10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpls.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmple_wide_n_u8(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpls.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmple_wide,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpls.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmple_wide_n_u16u10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpls.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_wide_n_u16(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpls.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmple_wide_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpls.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmple_wide_n_u32u10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpls.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmple_wide_n_u32(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmple_wide_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpls.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmple_wide,_n_u32,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmplt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmplt.c
index 9bd99e1766089..1ef15579ade27 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmplt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmplt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,343 +14,650 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmplt_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmplt_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmplt_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmplt_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmplt_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_s64u10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_s64(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpgt.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[DUP]], <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_u64u10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_u64(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmphi.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[DUP]], <vscale x 2 x i64> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmplt_wide_s8u10__SVBool_tu10__SVInt8_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmplt_wide_s8(svbool_t pg, svint8_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplt.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt_wide,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmplt_wide_s16u10__SVBool_tu11__SVInt16_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_wide_s16(svbool_t pg, svint16_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplt.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmplt_wide_s32u10__SVBool_tu11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_wide_s32(svbool_t pg, svint32_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplt.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplo.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmplt_wide_u8u10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplo.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmplt_wide_u8(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplo.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt_wide,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplo.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmplt_wide_u16u10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplo.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_wide_u16(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplo.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplo.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmplt_wide_u32u10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplo.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_wide_u32(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplo.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmplt_n_s8u10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmplt_n_s8(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[DUP]], <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_s16u10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_s16(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpgt.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[DUP]], <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_s32u10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_s32(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpgt.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[DUP]], <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmplt_n_u8u10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmplt_n_u8(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmphi.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[DUP]], <vscale x 16 x i8> %op1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_u16u10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_u16(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmphi.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[DUP]], <vscale x 8 x i16> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_u32u10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_u32(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmphi.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[DUP]], <vscale x 4 x i32> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmplt_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmplt_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpgt.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[DUP]], <vscale x 8 x half> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpgt.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[DUP]], <vscale x 4 x float> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmplt_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpgt.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[DUP]], <vscale x 2 x double> %op1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt,_n_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmplt_wide_n_s8u10__SVBool_tu10__SVInt8_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplt.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmplt_wide_n_s8(svbool_t pg, svint8_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplt.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt_wide,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmplt_wide_n_s16u10__SVBool_tu11__SVInt16_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplt.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_wide_n_s16(svbool_t pg, svint16_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplt.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmplt_wide_n_s32u10__SVBool_tu11__SVInt32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplt.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_wide_n_s32(svbool_t pg, svint32_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplt.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplo.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmplt_wide_n_u8u10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplo.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmplt_wide_n_u8(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmplo.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmplt_wide,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplo.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmplt_wide_n_u16u10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplo.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_wide_n_u16(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmplo.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmplt_wide_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplo.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmplt_wide_n_u32u10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplo.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmplt_wide_n_u32(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmplt_wide_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmplo.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmplt_wide,_n_u32,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpne.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpne.c
index 878864760f2b1..42bad6790261d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpne.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpne.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,284 +14,538 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmpne_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpne_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpne_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpne,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmpne_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpne_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpne,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_s64u10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_s64(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_u64u10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_u64(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_wide_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmpne_wide_s8u10__SVBool_tu10__SVInt8_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svcmpne_wide_s8(svbool_t pg, svint8_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_wide_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpne_wide,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_wide_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpne_wide_s16u10__SVBool_tu11__SVInt16_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_wide_s16(svbool_t pg, svint16_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_wide_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne_wide,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_wide_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcmpne_wide_s32u10__SVBool_tu11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_wide_s32(svbool_t pg, svint32_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_wide_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne_wide,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpne_n_s8u10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpne_n_s8(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpne,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_s16u10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_s16(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_s32u10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_s32(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmpne_n_u8u10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpne_n_u8(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpne,_n_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_u16u10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_u16(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_u32u10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_u32(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpne.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpne.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpne.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpne.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpne.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpne.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpne.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpne_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpne.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpne_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpne.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpne.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpne.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpne.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpne.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpne.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpne.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpne.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpne_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpne.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpne.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne,_n_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_wide_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmpne_wide_n_s8u10__SVBool_tu10__SVInt8_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svcmpne_wide_n_s8(svbool_t pg, svint8_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_wide_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcmpne_wide,_n_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_wide_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpne_wide_n_s16u10__SVBool_tu11__SVInt16_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_wide_n_s16(svbool_t pg, svint16_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_wide_n_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne_wide,_n_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpne_wide_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcmpne_wide_n_s32u10__SVBool_tu11__SVInt32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpne_wide_n_s32(svbool_t pg, svint32_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svcmpne_wide_n_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpne_wide,_n_s32,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpuo.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpuo.c
index 7531de36f1495..e8742c168fb0d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpuo.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cmpuo.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,65 +14,122 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmpuo_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpuo.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpuo_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpuo.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpuo_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svcmpuo_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpuo.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpuo,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpuo_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpuo.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpuo_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpuo.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpuo_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svcmpuo_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpuo.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpuo,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpuo_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpuo.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmpuo_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpuo.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svcmpuo_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svcmpuo_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpuo.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpuo,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpuo_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpuo.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpuo_n_f16u10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpuo.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpuo_n_f16(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svcmpuo_n_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.fcmpuo.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpuo,_n_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpuo_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpuo.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpuo_n_f32u10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpuo.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpuo_n_f32(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svcmpuo_n_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.fcmpuo.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpuo,_n_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svcmpuo_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpuo.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcmpuo_n_f64u10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpuo.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svcmpuo_n_f64(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svcmpuo_n_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.fcmpuo.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svcmpuo,_n_f64,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnot.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnot.c
index c0da8d1d4691f..1a277c0f6faa6 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnot.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnot.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,212 +14,398 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcnot_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnot_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcnot_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcnot_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svcnot_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcnot_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcnot_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcnot_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcnot_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcnot_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnot_u8_zu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnot_u8_z(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svcnot_u8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnot_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svcnot_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnot_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcnot_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnot_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcnot_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnot_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcnot_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcnot_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svcnot_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcnot_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcnot_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcnot_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcnot_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcnot_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnot_u8_mu11__SVUint8_tu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnot_u8_m(svuint8_t inactive, svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svcnot_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnot_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svcnot_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnot_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcnot_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnot_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcnot_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnot_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcnot_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcnot_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svcnot_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcnot_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcnot_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcnot_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcnot_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcnot_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnot_u8_xu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnot_u8_x(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svcnot_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnot.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnot_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svcnot_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnot.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnot_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcnot_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnot.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnot_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnot_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnot_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcnot_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnot.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnot,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt-bfloat.c
index 0bd8e036e5dda..b2d16157ebeba 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,28 +15,52 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcnt_bf16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnt_bf16_zu10__SVBool_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_bf16_z(svbool_t pg, svbfloat16_t op) {
-  // CHECK-LABEL: test_svcnt_bf16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcnt_bf16_z'}}
   return SVE_ACLE_FUNC(svcnt, _bf16, _z, )(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_bf16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnt_bf16_mu12__SVUint16_tu10__SVBool_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_bf16_m(svuint16_t inactive, svbool_t pg, svbfloat16_t op) {
-  // CHECK-LABEL: test_svcnt_bf16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcnt_bf16_m'}}
   return SVE_ACLE_FUNC(svcnt, _bf16, _m, )(inactive, pg, op);
 }
+// CHECK-LABEL: @test_svcnt_bf16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnt_bf16_xu10__SVBool_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_bf16_x(svbool_t pg, svbfloat16_t op) {
-  // CHECK-LABEL: test_svcnt_bf16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8bf16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcnt_bf16_x'}}
   return SVE_ACLE_FUNC(svcnt, _bf16, _x, )(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt.c
index 59af5391ad0c4..f9676c630c195 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,293 +14,551 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcnt_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnt_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnt_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcnt_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcnt_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcnt_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcnt_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnt_u8_zu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnt_u8_z(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svcnt_u8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svcnt_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcnt_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcnt_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcnt_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcnt_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcnt_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnt_s8_mu11__SVUint8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnt_s8_m(svuint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcnt_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s16_mu12__SVUint16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_s16_m(svuint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcnt_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s32_mu12__SVUint32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_s32_m(svuint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcnt_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s64_mu12__SVUint64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_s64_m(svuint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcnt_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnt_u8_mu11__SVUint8_tu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnt_u8_m(svuint8_t inactive, svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svcnt_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svcnt_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcnt_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcnt_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f16_mu12__SVUint16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_f16_m(svuint16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcnt_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f32_mu12__SVUint32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_f32_m(svuint32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcnt_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f64_mu12__SVUint64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_f64_m(svuint64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcnt_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnt_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnt_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svcnt_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svcnt_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcnt_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcnt_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnt_u8_xu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcnt_u8_x(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svcnt_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cnt.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svcnt_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcnt_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcnt_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_u64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcnt_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcnt_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cnt.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcnt_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcnt_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cnt.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcnt_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcnt_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcnt_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcnt_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cnt.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcnt,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntb.c
index 9e4326e312c54..a5991a5a0151c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntb.c
@@ -1,142 +1,262 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svcntb(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z11test_svcntbv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcntb()
 {
-  // CHECK-LABEL: test_svcntb
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 4
-  // CHECK: ret i64 %[[RET]]
   return svcntb();
 }
 
+// CHECK-LABEL: @test_svcntb_pat(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 0)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcntb_patv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 0)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat()
 {
-  // CHECK-LABEL: test_svcntb_pat
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 0)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_POW2);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 1
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_1v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 1
+//
 uint64_t test_svcntb_pat_1()
 {
-  // CHECK-LABEL: test_svcntb_pat_1
-  // CHECK: ret i64 1
   return svcntb_pat(SV_VL1);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 2
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_2v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 2
+//
 uint64_t test_svcntb_pat_2()
 {
-  // CHECK-LABEL: test_svcntb_pat_2
-  // CHECK: ret i64 2
   return svcntb_pat(SV_VL2);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 3
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_3v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 3
+//
 uint64_t test_svcntb_pat_3()
 {
-  // CHECK-LABEL: test_svcntb_pat_3
-  // CHECK: ret i64 3
   return svcntb_pat(SV_VL3);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 4
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_4v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 4
+//
 uint64_t test_svcntb_pat_4()
 {
-  // CHECK-LABEL: test_svcntb_pat_4
-  // CHECK: ret i64 4
   return svcntb_pat(SV_VL4);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 5
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_5v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 5
+//
 uint64_t test_svcntb_pat_5()
 {
-  // CHECK-LABEL: test_svcntb_pat_5
-  // CHECK: ret i64 5
   return svcntb_pat(SV_VL5);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 6
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_6v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 6
+//
 uint64_t test_svcntb_pat_6()
 {
-  // CHECK-LABEL: test_svcntb_pat_6
-  // CHECK: ret i64 6
   return svcntb_pat(SV_VL6);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 7
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_7v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 7
+//
 uint64_t test_svcntb_pat_7()
 {
-  // CHECK-LABEL: test_svcntb_pat_7
-  // CHECK: ret i64 7
   return svcntb_pat(SV_VL7);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 8
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 8
+//
 uint64_t test_svcntb_pat_8()
 {
-  // CHECK-LABEL: test_svcntb_pat_8
-  // CHECK: ret i64 8
   return svcntb_pat(SV_VL8);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 16
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntb_pat_9v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 16
+//
 uint64_t test_svcntb_pat_9()
 {
-  // CHECK-LABEL: test_svcntb_pat_9
-  // CHECK: ret i64 16
   return svcntb_pat(SV_VL16);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 10)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_10v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 10)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat_10()
 {
-  // CHECK-LABEL: test_svcntb_pat_10
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 10)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_VL32);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 11)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_11v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 11)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat_11()
 {
-  // CHECK-LABEL: test_svcntb_pat_11
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 11)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_VL64);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_12(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 12)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_12v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 12)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat_12()
 {
-  // CHECK-LABEL: test_svcntb_pat_12
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 12)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_VL128);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_13(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 13)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_13v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 13)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat_13()
 {
-  // CHECK-LABEL: test_svcntb_pat_13
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 13)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_VL256);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_14(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 29)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_14v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 29)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat_14()
 {
-  // CHECK-LABEL: test_svcntb_pat_14
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 29)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_MUL4);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_15(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 30)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_15v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntb(i32 30)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntb_pat_15()
 {
-  // CHECK-LABEL: test_svcntb_pat_15
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntb(i32 30)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntb_pat(SV_MUL3);
 }
 
+// CHECK-LABEL: @test_svcntb_pat_16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntb_pat_16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcntb_pat_16()
 {
-  // CHECK-LABEL: test_svcntb_pat_16
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 4
-  // CHECK: ret i64 %[[RET]]
   return svcntb_pat(SV_ALL);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntd.c
index 9880968bae9b0..eb3e848a7b532 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntd.c
@@ -1,149 +1,276 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svcntd(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z11test_svcntdv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcntd()
 {
-  // CHECK-LABEL: test_svcntd
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 1
-  // CHECK: ret i64 %[[RET]]
   return svcntd();
 }
 
+// CHECK-LABEL: @test_svcntd_pat(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 0)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcntd_patv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 0)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat()
 {
-  // CHECK-LABEL: test_svcntd_pat
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 0)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_POW2);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 1
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_1v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 1
+//
 uint64_t test_svcntd_pat_1()
 {
-  // CHECK-LABEL: test_svcntd_pat_1
-  // CHECK: ret i64 1
   return svcntd_pat(SV_VL1);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 2
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_2v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 2
+//
 uint64_t test_svcntd_pat_2()
 {
-  // CHECK-LABEL: test_svcntd_pat_2
-  // CHECK: ret i64 2
   return svcntd_pat(SV_VL2);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 3)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_3v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 3)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_3()
 {
-  // CHECK-LABEL: test_svcntd_pat_3
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 3)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL3);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 4)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_4v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 4)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_4()
 {
-  // CHECK-LABEL: test_svcntd_pat_4
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 4)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL4);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 5)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_5v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 5)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_5()
 {
-  // CHECK-LABEL: test_svcntd_pat_5
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 5)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL5);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 6)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_6v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 6)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_6()
 {
-  // CHECK-LABEL: test_svcntd_pat_6
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 6)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL6);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 7)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_7v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 7)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_7()
 {
-  // CHECK-LABEL: test_svcntd_pat_7
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 7)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL7);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 8)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 8)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_8()
 {
-  // CHECK-LABEL: test_svcntd_pat_8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 8)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL8);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 9)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntd_pat_9v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 9)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_9()
 {
-  // CHECK-LABEL: test_svcntd_pat_9
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 9)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL16);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 10)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_10v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 10)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_10()
 {
-  // CHECK-LABEL: test_svcntd_pat_10
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 10)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL32);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 11)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_11v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 11)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_11()
 {
-  // CHECK-LABEL: test_svcntd_pat_11
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 11)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL64);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_12(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 12)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_12v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 12)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_12()
 {
-  // CHECK-LABEL: test_svcntd_pat_12
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 12)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL128);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_13(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 13)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_13v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 13)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_13()
 {
-  // CHECK-LABEL: test_svcntd_pat_13
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 13)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_VL256);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_14(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 29)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_14v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 29)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_14()
 {
-  // CHECK-LABEL: test_svcntd_pat_14
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 29)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_MUL4);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_15(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 30)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_15v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntd(i32 30)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntd_pat_15()
 {
-  // CHECK-LABEL: test_svcntd_pat_15
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntd(i32 30)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntd_pat(SV_MUL3);
 }
 
+// CHECK-LABEL: @test_svcntd_pat_16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntd_pat_16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcntd_pat_16()
 {
-  // CHECK-LABEL: test_svcntd_pat_16
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 1
-  // CHECK: ret i64 %[[RET]]
   return svcntd_pat(SV_ALL);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnth.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnth.c
index 4c3c03d2ca302..50ca8f525387b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnth.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cnth.c
@@ -1,143 +1,264 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svcnth(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z11test_svcnthv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcnth()
 {
-  // CHECK-LABEL: test_svcnth
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 3
-  // CHECK: ret i64 %[[RET]]
   return svcnth();
 }
 
+// CHECK-LABEL: @test_svcnth_pat(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 0)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcnth_patv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 0)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat()
 {
-  // CHECK-LABEL: test_svcnth_pat
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 0)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_POW2);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 1
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_1v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 1
+//
 uint64_t test_svcnth_pat_1()
 {
-  // CHECK-LABEL: test_svcnth_pat_1
-  // CHECK: ret i64 1
   return svcnth_pat(SV_VL1);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 2
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_2v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 2
+//
 uint64_t test_svcnth_pat_2()
 {
-  // CHECK-LABEL: test_svcnth_pat_2
-  // CHECK: ret i64 2
   return svcnth_pat(SV_VL2);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 3
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_3v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 3
+//
 uint64_t test_svcnth_pat_3()
 {
-  // CHECK-LABEL: test_svcnth_pat_3
-  // CHECK: ret i64 3
   return svcnth_pat(SV_VL3);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 4
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_4v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 4
+//
 uint64_t test_svcnth_pat_4()
 {
-  // CHECK-LABEL: test_svcnth_pat_4
-  // CHECK: ret i64 4
   return svcnth_pat(SV_VL4);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 5
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_5v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 5
+//
 uint64_t test_svcnth_pat_5()
 {
-  // CHECK-LABEL: test_svcnth_pat_5
-  // CHECK: ret i64 5
   return svcnth_pat(SV_VL5);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 6
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_6v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 6
+//
 uint64_t test_svcnth_pat_6()
 {
-  // CHECK-LABEL: test_svcnth_pat_6
-  // CHECK: ret i64 6
   return svcnth_pat(SV_VL6);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 7
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_7v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 7
+//
 uint64_t test_svcnth_pat_7()
 {
-  // CHECK-LABEL: test_svcnth_pat_7
-  // CHECK: ret i64 7
   return svcnth_pat(SV_VL7);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 8
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 8
+//
 uint64_t test_svcnth_pat_8()
 {
-  // CHECK-LABEL: test_svcnth_pat_8
-  // CHECK: ret i64 8
   return svcnth_pat(SV_VL8);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 9)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcnth_pat_9v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 9)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_9()
 {
-  // CHECK-LABEL: test_svcnth_pat_9
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 9)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_VL16);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 10)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_10v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 10)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_10()
 {
-  // CHECK-LABEL: test_svcnth_pat_10
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 10)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_VL32);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 11)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_11v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 11)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_11()
 {
-  // CHECK-LABEL: test_svcnth_pat_11
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 11)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_VL64);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_12(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 12)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_12v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 12)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_12()
 {
-  // CHECK-LABEL: test_svcnth_pat_12
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 12)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_VL128);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_13(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 13)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_13v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 13)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_13()
 {
-  // CHECK-LABEL: test_svcnth_pat_13
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 13)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_VL256);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_14(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 29)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_14v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 29)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_14()
 {
-  // CHECK-LABEL: test_svcnth_pat_14
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 29)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_MUL4);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_15(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 30)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_15v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cnth(i32 30)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcnth_pat_15()
 {
-  // CHECK-LABEL: test_svcnth_pat_15
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cnth(i32 30)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcnth_pat(SV_MUL3);
 }
 
+// CHECK-LABEL: @test_svcnth_pat_16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcnth_pat_16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcnth_pat_16()
 {
-  // CHECK-LABEL: test_svcnth_pat_16
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 3
-  // CHECK: ret i64 %[[RET]]
   return svcnth_pat(SV_ALL);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntp.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntp.c
index 1fe39d45d21ce..4850cecf31e62 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntp.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntp.c
@@ -1,43 +1,78 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svcntp_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svcntp_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntp_b8(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svcntp_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntp.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntp_b8(pg, op);
 }
 
+// CHECK-LABEL: @test_svcntp_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret i64 [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcntp_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP2]]
+//
 uint64_t test_svcntp_b16(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svcntp_b16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntp.nxv8i1(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i1> %[[OP]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntp_b16(pg, op);
 }
 
+// CHECK-LABEL: @test_svcntp_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret i64 [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcntp_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP2]]
+//
 uint64_t test_svcntp_b32(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svcntp_b32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntp.nxv4i1(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i1> %[[OP]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntp_b32(pg, op);
 }
 
+// CHECK-LABEL: @test_svcntp_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret i64 [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcntp_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call i64 @llvm.aarch64.sve.cntp.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP2]]
+//
 uint64_t test_svcntp_b64(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svcntp_b64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntp.nxv2i1(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i1> %[[OP]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntp_b64(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntw.c
index a57c5d2bdb151..4c12fd8ee7b4a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cntw.c
@@ -1,147 +1,272 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svcntw(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z11test_svcntwv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcntw()
 {
-  // CHECK-LABEL: test_svcntw
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 2
-  // CHECK: ret i64 %[[RET]]
   return svcntw();
 }
 
+// CHECK-LABEL: @test_svcntw_pat(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 0)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcntw_patv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 0)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat()
 {
-  // CHECK-LABEL: test_svcntw_pat
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 0)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_POW2);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 1
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_1v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 1
+//
 uint64_t test_svcntw_pat_1()
 {
-  // CHECK-LABEL: test_svcntw_pat_1
-  // CHECK: ret i64 1
   return svcntw_pat(SV_VL1);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 2
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_2v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 2
+//
 uint64_t test_svcntw_pat_2()
 {
-  // CHECK-LABEL: test_svcntw_pat_2
-  // CHECK: ret i64 2
   return svcntw_pat(SV_VL2);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 3
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_3v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 3
+//
 uint64_t test_svcntw_pat_3()
 {
-  // CHECK-LABEL: test_svcntw_pat_3
-  // CHECK: ret i64 3
   return svcntw_pat(SV_VL3);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret i64 4
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_4v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret i64 4
+//
 uint64_t test_svcntw_pat_4()
 {
-  // CHECK-LABEL: test_svcntw_pat_4
-  // CHECK: ret i64 4
   return svcntw_pat(SV_VL4);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 5)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_5v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 5)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_5()
 {
-  // CHECK-LABEL: test_svcntw_pat_5
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 5)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL5);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 6)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_6v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 6)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_6()
 {
-  // CHECK-LABEL: test_svcntw_pat_6
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 6)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL6);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 7)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_7v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 7)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_7()
 {
-  // CHECK-LABEL: test_svcntw_pat_7
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 7)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL7);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 8)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 8)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_8()
 {
-  // CHECK-LABEL: test_svcntw_pat_8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 8)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL8);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 9)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcntw_pat_9v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 9)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_9()
 {
-  // CHECK-LABEL: test_svcntw_pat_9
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 9)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL16);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 10)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_10v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 10)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_10()
 {
-  // CHECK-LABEL: test_svcntw_pat_10
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 10)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL32);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 11)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_11v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 11)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_11()
 {
-  // CHECK-LABEL: test_svcntw_pat_11
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 11)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL64);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_12(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 12)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_12v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 12)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_12()
 {
-  // CHECK-LABEL: test_svcntw_pat_12
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 12)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL128);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_13(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 13)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_13v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 13)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_13()
 {
-  // CHECK-LABEL: test_svcntw_pat_13
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 13)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_VL256);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_14(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 29)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_14v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 29)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_14()
 {
-  // CHECK-LABEL: test_svcntw_pat_14
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 29)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_MUL4);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_15(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 30)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_15v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.cntw(i32 30)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svcntw_pat_15()
 {
-  // CHECK-LABEL: test_svcntw_pat_15
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.cntw(i32 30)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return svcntw_pat(SV_MUL3);
 }
 
+// CHECK-LABEL: @test_svcntw_pat_16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcntw_pat_16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svcntw_pat_16()
 {
-  // CHECK-LABEL: test_svcntw_pat_16
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK-NEXT: %[[RET:.*]] = shl i64 %[[INTRINSIC]], 2
-  // CHECK: ret i64 %[[RET]]
   return svcntw_pat(SV_ALL);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_compact.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_compact.c
index 89d2d02b79929..69f3ff9c782a1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_compact.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_compact.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,56 +14,104 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcompact_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.compact.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcompact_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.compact.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcompact_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcompact_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.compact.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcompact,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcompact_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.compact.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcompact_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.compact.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcompact_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcompact_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.compact.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcompact,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcompact_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.compact.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcompact_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.compact.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcompact_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcompact_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.compact.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcompact,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcompact_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.compact.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcompact_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.compact.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcompact_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcompact_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.compact.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcompact,_u64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcompact_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.compact.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcompact_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.compact.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcompact_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcompact_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.compact.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcompact,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcompact_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.compact.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcompact_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.compact.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcompact_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcompact_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.compact.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcompact,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
index f4b81a671d093..d41179bb2ac78 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcreate2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.create2.nxv16bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcreate2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.create2.nxv16bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
 svbfloat16x2_t test_svcreate2_bf16(svbfloat16_t x0, svbfloat16_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_bf16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.create2.nxv16bf16.nxv8bf16(<vscale x 8 x bfloat> %x0, <vscale x 8 x bfloat> %x1)
-  // CHECK-NEXT: ret <vscale x 16 x bfloat> %[[CREATE]]
   // expected-warning at +1 {{implicit declaration of function 'svcreate2_bf16'}}
   return SVE_ACLE_FUNC(svcreate2,_bf16,,)(x0, x1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
index db2a6c5521143..de04db3f56097 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create2.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,90 +13,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcreate2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcreate2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
 svint8x2_t test_svcreate2_s8(svint8_t x0, svint8_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_s8
-  // CHECK: %[[CREATE:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> %x0, <vscale x 16 x i8> %x1)
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_s8,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
 svint16x2_t test_svcreate2_s16(svint16_t x0, svint16_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_s16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> %x0, <vscale x 8 x i16> %x1)
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_s16,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
 svint32x2_t test_svcreate2_s32(svint32_t x0, svint32_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_s32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> %x0, <vscale x 4 x i32> %x1)
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_s32,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
 svint64x2_t test_svcreate2_s64(svint64_t x0, svint64_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_s64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> %x0, <vscale x 2 x i64> %x1)
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_s64,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcreate2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
 svuint8x2_t test_svcreate2_u8(svuint8_t x0, svuint8_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_u8
-  // CHECK: %[[CREATE:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.create2.nxv32i8.nxv16i8(<vscale x 16 x i8> %x0, <vscale x 16 x i8> %x1)
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_u8,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
 svuint16x2_t test_svcreate2_u16(svuint16_t x0, svuint16_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_u16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.create2.nxv16i16.nxv8i16(<vscale x 8 x i16> %x0, <vscale x 8 x i16> %x1)
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_u16,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
 svuint32x2_t test_svcreate2_u32(svuint32_t x0, svuint32_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_u32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.create2.nxv8i32.nxv4i32(<vscale x 4 x i32> %x0, <vscale x 4 x i32> %x1)
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_u32,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
 svuint64x2_t test_svcreate2_u64(svuint64_t x0, svuint64_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_u64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.create2.nxv4i64.nxv2i64(<vscale x 2 x i64> %x0, <vscale x 2 x i64> %x1)
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_u64,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.tuple.create2.nxv16f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.tuple.create2.nxv16f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
+//
 svfloat16x2_t test_svcreate2_f16(svfloat16_t x0, svfloat16_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_f16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.tuple.create2.nxv16f16.nxv8f16(<vscale x 8 x half> %x0, <vscale x 8 x half> %x1)
-  // CHECK-NEXT: ret <vscale x 16 x half> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_f16,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.tuple.create2.nxv8f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.tuple.create2.nxv8f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
 svfloat32x2_t test_svcreate2_f32(svfloat32_t x0, svfloat32_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_f32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.tuple.create2.nxv8f32.nxv4f32(<vscale x 4 x float> %x0, <vscale x 4 x float> %x1)
-  // CHECK-NEXT: ret <vscale x 8 x float> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_f32,,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svcreate2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.tuple.create2.nxv4f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.tuple.create2.nxv4f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
+//
 svfloat64x2_t test_svcreate2_f64(svfloat64_t x0, svfloat64_t x1)
 {
-  // CHECK-LABEL: test_svcreate2_f64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.tuple.create2.nxv4f64.nxv2f64(<vscale x 2 x double> %x0, <vscale x 2 x double> %x1)
-  // CHECK-NEXT: ret <vscale x 4 x double> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate2,_f64,,)(x0, x1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
index 7ec4814fd111a..eefdf1c07664d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcreate3_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.create3.nxv24bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcreate3_bf16u14__SVBFloat16_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.create3.nxv24bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
 svbfloat16x3_t test_svcreate3_bf16(svbfloat16_t x0, svbfloat16_t x1, svbfloat16_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_bf16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.create3.nxv24bf16.nxv8bf16(<vscale x 8 x bfloat> %x0, <vscale x 8 x bfloat> %x1, <vscale x 8 x bfloat> %x2)
-  // CHECK-NEXT: ret <vscale x 24 x bfloat> %[[CREATE]]
   // expected-warning at +1 {{implicit declaration of function 'svcreate3_bf16'}}
   return SVE_ACLE_FUNC(svcreate3,_bf16,,)(x0, x1, x2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
index 37ada8a187378..2148bda5dac52 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create3.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,90 +13,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcreate3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcreate3_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
 svint8x3_t test_svcreate3_s8(svint8_t x0, svint8_t x1, svint8_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_s8
-  // CHECK: %[[CREATE:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> %x0, <vscale x 16 x i8> %x1, <vscale x 16 x i8> %x2)
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_s8,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
 svint16x3_t test_svcreate3_s16(svint16_t x0, svint16_t x1, svint16_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_s16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> %x0, <vscale x 8 x i16> %x1, <vscale x 8 x i16> %x2)
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_s16,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
 svint32x3_t test_svcreate3_s32(svint32_t x0, svint32_t x1, svint32_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_s32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> %x0, <vscale x 4 x i32> %x1, <vscale x 4 x i32> %x2)
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_s32,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
 svint64x3_t test_svcreate3_s64(svint64_t x0, svint64_t x1, svint64_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_s64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> %x0, <vscale x 2 x i64> %x1, <vscale x 2 x i64> %x2)
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_s64,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcreate3_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
 svuint8x3_t test_svcreate3_u8(svuint8_t x0, svuint8_t x1, svuint8_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_u8
-  // CHECK: %[[CREATE:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.create3.nxv48i8.nxv16i8(<vscale x 16 x i8> %x0, <vscale x 16 x i8> %x1, <vscale x 16 x i8> %x2)
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_u8,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
 svuint16x3_t test_svcreate3_u16(svuint16_t x0, svuint16_t x1, svuint16_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_u16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.create3.nxv24i16.nxv8i16(<vscale x 8 x i16> %x0, <vscale x 8 x i16> %x1, <vscale x 8 x i16> %x2)
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_u16,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
 svuint32x3_t test_svcreate3_u32(svuint32_t x0, svuint32_t x1, svuint32_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_u32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.create3.nxv12i32.nxv4i32(<vscale x 4 x i32> %x0, <vscale x 4 x i32> %x1, <vscale x 4 x i32> %x2)
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_u32,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
 svuint64x3_t test_svcreate3_u64(svuint64_t x0, svuint64_t x1, svuint64_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_u64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.create3.nxv6i64.nxv2i64(<vscale x 2 x i64> %x0, <vscale x 2 x i64> %x1, <vscale x 2 x i64> %x2)
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_u64,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.tuple.create3.nxv24f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.tuple.create3.nxv24f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
+//
 svfloat16x3_t test_svcreate3_f16(svfloat16_t x0, svfloat16_t x1, svfloat16_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_f16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.tuple.create3.nxv24f16.nxv8f16(<vscale x 8 x half> %x0, <vscale x 8 x half> %x1, <vscale x 8 x half> %x2)
-  // CHECK-NEXT: ret <vscale x 24 x half> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_f16,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.tuple.create3.nxv12f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.tuple.create3.nxv12f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
+//
 svfloat32x3_t test_svcreate3_f32(svfloat32_t x0, svfloat32_t x1, svfloat32_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_f32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.tuple.create3.nxv12f32.nxv4f32(<vscale x 4 x float> %x0, <vscale x 4 x float> %x1, <vscale x 4 x float> %x2)
-  // CHECK-NEXT: ret <vscale x 12 x float> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_f32,,)(x0, x1, x2);
 }
 
+// CHECK-LABEL: @test_svcreate3_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.tuple.create3.nxv6f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate3_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.tuple.create3.nxv6f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
+//
 svfloat64x3_t test_svcreate3_f64(svfloat64_t x0, svfloat64_t x1, svfloat64_t x2)
 {
-  // CHECK-LABEL: test_svcreate3_f64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.tuple.create3.nxv6f64.nxv2f64(<vscale x 2 x double> %x0, <vscale x 2 x double> %x1, <vscale x 2 x double> %x2)
-  // CHECK-NEXT: ret <vscale x 6 x double> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate3,_f64,,)(x0, x1, x2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
index 17723b9419a7b..97636facc9a32 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcreate4_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.create4.nxv32bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]], <vscale x 8 x bfloat> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcreate4_bf16u14__SVBFloat16_tu14__SVBFloat16_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.create4.nxv32bf16.nxv8bf16(<vscale x 8 x bfloat> [[X0:%.*]], <vscale x 8 x bfloat> [[X1:%.*]], <vscale x 8 x bfloat> [[X2:%.*]], <vscale x 8 x bfloat> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
 svbfloat16x4_t test_svcreate4_bf16(svbfloat16_t x0, svbfloat16_t x1, svbfloat16_t x2, svbfloat16_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_bf16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.create4.nxv32bf16.nxv8bf16(<vscale x 8 x bfloat> %x0, <vscale x 8 x bfloat> %x1, <vscale x 8 x bfloat> %x2, <vscale x 8 x bfloat> %x4)
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[CREATE]]
   // expected-warning at +1 {{implicit declaration of function 'svcreate4_bf16'}}
   return SVE_ACLE_FUNC(svcreate4,_bf16,,)(x0, x1, x2, x4);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
index f30bbb880f83f..ee4053a29ab8b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_create4.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,90 +13,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcreate4_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcreate4_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
 svint8x4_t test_svcreate4_s8(svint8_t x0, svint8_t x1, svint8_t x2, svint8_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_s8
-  // CHECK: %[[CREATE:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> %x0, <vscale x 16 x i8> %x1, <vscale x 16 x i8> %x2, <vscale x 16 x i8> %x4)
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_s8,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
 svint16x4_t test_svcreate4_s16(svint16_t x0, svint16_t x1, svint16_t x2, svint16_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_s16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> %x0, <vscale x 8 x i16> %x1, <vscale x 8 x i16> %x2, <vscale x 8 x i16> %x4)
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_s16,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
 svint32x4_t test_svcreate4_s32(svint32_t x0, svint32_t x1, svint32_t x2, svint32_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_s32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> %x0, <vscale x 4 x i32> %x1, <vscale x 4 x i32> %x2, <vscale x 4 x i32> %x4)
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_s32,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
 svint64x4_t test_svcreate4_s64(svint64_t x0, svint64_t x1, svint64_t x2, svint64_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_s64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> %x0, <vscale x 2 x i64> %x1, <vscale x 2 x i64> %x2, <vscale x 2 x i64> %x4)
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_s64,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcreate4_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> [[X0:%.*]], <vscale x 16 x i8> [[X1:%.*]], <vscale x 16 x i8> [[X2:%.*]], <vscale x 16 x i8> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
 svuint8x4_t test_svcreate4_u8(svuint8_t x0, svuint8_t x1, svuint8_t x2, svuint8_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_u8
-  // CHECK: %[[CREATE:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.create4.nxv64i8.nxv16i8(<vscale x 16 x i8> %x0, <vscale x 16 x i8> %x1, <vscale x 16 x i8> %x2, <vscale x 16 x i8> %x4)
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_u8,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> [[X0:%.*]], <vscale x 8 x i16> [[X1:%.*]], <vscale x 8 x i16> [[X2:%.*]], <vscale x 8 x i16> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
 svuint16x4_t test_svcreate4_u16(svuint16_t x0, svuint16_t x1, svuint16_t x2, svuint16_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_u16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.create4.nxv32i16.nxv8i16(<vscale x 8 x i16> %x0, <vscale x 8 x i16> %x1, <vscale x 8 x i16> %x2, <vscale x 8 x i16> %x4)
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_u16,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> [[X0:%.*]], <vscale x 4 x i32> [[X1:%.*]], <vscale x 4 x i32> [[X2:%.*]], <vscale x 4 x i32> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
 svuint32x4_t test_svcreate4_u32(svuint32_t x0, svuint32_t x1, svuint32_t x2, svuint32_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_u32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.create4.nxv16i32.nxv4i32(<vscale x 4 x i32> %x0, <vscale x 4 x i32> %x1, <vscale x 4 x i32> %x2, <vscale x 4 x i32> %x4)
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_u32,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> [[X0:%.*]], <vscale x 2 x i64> [[X1:%.*]], <vscale x 2 x i64> [[X2:%.*]], <vscale x 2 x i64> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
 svuint64x4_t test_svcreate4_u64(svuint64_t x0, svuint64_t x1, svuint64_t x2, svuint64_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_u64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.create4.nxv8i64.nxv2i64(<vscale x 2 x i64> %x0, <vscale x 2 x i64> %x1, <vscale x 2 x i64> %x2, <vscale x 2 x i64> %x4)
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_u64,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.tuple.create4.nxv32f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]], <vscale x 8 x half> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.tuple.create4.nxv32f16.nxv8f16(<vscale x 8 x half> [[X0:%.*]], <vscale x 8 x half> [[X1:%.*]], <vscale x 8 x half> [[X2:%.*]], <vscale x 8 x half> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
+//
 svfloat16x4_t test_svcreate4_f16(svfloat16_t x0, svfloat16_t x1, svfloat16_t x2, svfloat16_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_f16
-  // CHECK: %[[CREATE:.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.tuple.create4.nxv32f16.nxv8f16(<vscale x 8 x half> %x0, <vscale x 8 x half> %x1, <vscale x 8 x half> %x2, <vscale x 8 x half> %x4)
-  // CHECK-NEXT: ret <vscale x 32 x half> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_f16,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.tuple.create4.nxv16f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]], <vscale x 4 x float> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.tuple.create4.nxv16f32.nxv4f32(<vscale x 4 x float> [[X0:%.*]], <vscale x 4 x float> [[X1:%.*]], <vscale x 4 x float> [[X2:%.*]], <vscale x 4 x float> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
 svfloat32x4_t test_svcreate4_f32(svfloat32_t x0, svfloat32_t x1, svfloat32_t x2, svfloat32_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_f32
-  // CHECK: %[[CREATE:.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.tuple.create4.nxv16f32.nxv4f32(<vscale x 4 x float> %x0, <vscale x 4 x float> %x1, <vscale x 4 x float> %x2, <vscale x 4 x float> %x4)
-  // CHECK-NEXT: ret <vscale x 16 x float> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_f32,,)(x0, x1, x2, x4);
 }
 
+// CHECK-LABEL: @test_svcreate4_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.tuple.create4.nxv8f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]], <vscale x 2 x double> [[X4:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svcreate4_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.tuple.create4.nxv8f64.nxv2f64(<vscale x 2 x double> [[X0:%.*]], <vscale x 2 x double> [[X1:%.*]], <vscale x 2 x double> [[X2:%.*]], <vscale x 2 x double> [[X4:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
+//
 svfloat64x4_t test_svcreate4_f64(svfloat64_t x0, svfloat64_t x1, svfloat64_t x2, svfloat64_t x4)
 {
-  // CHECK-LABEL: test_svcreate4_f64
-  // CHECK: %[[CREATE:.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.tuple.create4.nxv8f64.nxv2f64(<vscale x 2 x double> %x0, <vscale x 2 x double> %x1, <vscale x 2 x double> %x2, <vscale x 2 x double> %x4)
-  // CHECK-NEXT: ret <vscale x 8 x double> %[[CREATE]]
   return SVE_ACLE_FUNC(svcreate4,_f64,,)(x0, x1, x2, x4);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt-bfloat.c
index f7f397aba059f..fff5ddc4fb17b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,26 +13,50 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvt_bf16_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcvt_bf16_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svcvt_bf16_f32_x(svbool_t pg, svfloat32_t op) {
-  // CHECK-LABEL: test_svcvt_bf16_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_bf16, _f32, _x, )(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_bf16_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcvt_bf16_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svcvt_bf16_f32_z(svbool_t pg, svfloat32_t op) {
-  // CHECK-LABEL: test_svcvt_bf16_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_bf16, _f32, _z, )(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_bf16_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcvt_bf16_f32_mu14__SVBFloat16_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svcvt_bf16_f32_m(svbfloat16_t inactive, svbool_t pg, svfloat32_t op) {
-  // CHECK-LABEL: test_svcvt_bf16_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvt.bf16f32(<vscale x 8 x bfloat> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_bf16, _f32, _m, )(inactive, pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt.c
index 3f4cb79890144..207897876281e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,866 +14,1634 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvt_s16_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s16_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svcvt_s16_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s16_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s16,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s16_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s16_f16_mu11__SVInt16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svcvt_s16_f16_m(svint16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s16_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s16,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s16_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s16_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svcvt_s16_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s16_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzs.nxv8i16.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s16,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u16_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u16_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcvt_u16_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u16_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u16,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u16_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u16_f16_mu12__SVUint16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcvt_u16_f16_m(svuint16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u16_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u16,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u16_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u16_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svcvt_u16_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u16_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.fcvtzu.nxv8i16.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u16,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f16_mu11__SVInt32_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f16_m(svint32_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f32_mu11__SVInt32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f32_m(svint32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f64_mu11__SVInt32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f64_m(svint32_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f16(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.nxv4i32.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s32_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s32_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svcvt_s32_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_s32_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzs.i32f64(<vscale x 4 x i32> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s32,_f64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f16_mu11__SVInt64_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f16_m(svint64_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f32_mu11__SVInt64_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f32_m(svint64_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f64_mu11__SVInt64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f64_m(svint64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f16(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.i64f32(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_s64_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_s64_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svcvt_s64_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_s64_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzs.nxv2i64.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_s64,_f64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f16_mu12__SVUint32_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f16_m(svuint32_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f32_mu12__SVUint32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f32_m(svuint32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f64_mu12__SVUint32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f64_m(svuint32_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f16(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.nxv4i32.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u32_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u32_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svcvt_u32_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_u32_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.fcvtzu.i32f64(<vscale x 4 x i32> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u32,_f64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f16_mu12__SVUint64_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f16_m(svuint64_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f32_mu12__SVUint64_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f32_m(svuint64_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f64_mu12__SVUint64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f64_m(svuint64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f16(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.i64f32(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_u64_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_u64_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svcvt_u64_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_u64_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.fcvtzu.nxv2i64.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_u64,_f64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_s32_mu13__SVFloat16_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_s32_m(svfloat16_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_s32_mu13__SVFloat32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_s32_m(svfloat32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_s32_mu13__SVFloat64_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_s32_m(svfloat64_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i32(<vscale x 8 x half> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.nxv4f32.nxv4i32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.f64i32(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_s64_mu13__SVFloat16_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_s64_m(svfloat16_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_s64_mu13__SVFloat32_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_s64_m(svfloat32_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_s64_mu13__SVFloat64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_s64_m(svfloat64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.scvtf.f16i64(<vscale x 8 x half> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.scvtf.f32i64(<vscale x 4 x float> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.scvtf.nxv2f64.nxv2i64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_u32_mu13__SVFloat16_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_u32_m(svfloat16_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_u32_mu13__SVFloat32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_u32_m(svfloat32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_u32_mu13__SVFloat64_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_u32_m(svfloat64_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i32(<vscale x 8 x half> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.nxv4f32.nxv4i32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.f64i32(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_u64_mu13__SVFloat16_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_u64_m(svfloat16_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_u64_mu13__SVFloat32_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_u64_m(svfloat32_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_u64_mu13__SVFloat64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_u64_m(svfloat64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ucvtf.f16i64(<vscale x 8 x half> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_u64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ucvtf.f32i64(<vscale x 4 x float> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_u64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ucvtf.nxv2f64.nxv2i64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_u64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_f16_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_f16_m(svfloat32_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_f16_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_f16_m(svfloat64_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f16(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f16(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_f32_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_f32_m(svfloat64_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f64_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f64_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvt_f64_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f64_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvt.f64f32(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f64,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_f32_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_f32_m(svfloat16_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_f64_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_f64_m(svfloat16_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f32(<vscale x 8 x half> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f16_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f16_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvt_f16_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f16_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvt.f16f64(<vscale x 8 x half> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f16,_f64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_f64_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_f64_m(svfloat32_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvt_f32_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcvt_f32_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvt_f32_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvt_f32_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvt.f32f64(<vscale x 4 x float> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvt_f32,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvtnt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvtnt.c
index 8dee5e5e4f8e2..e9a032119819e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvtnt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_cvtnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,18 +13,34 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvtnt_bf16_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvtnt.bf16f32(<vscale x 8 x bfloat> [[EVEN:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcvtnt_bf16_f32_xu14__SVBFloat16_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvtnt.bf16f32(<vscale x 8 x bfloat> [[EVEN:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svcvtnt_bf16_f32_x(svbfloat16_t even, svbool_t pg, svfloat32_t op) {
-  // CHECK-LABEL: test_svcvtnt_bf16_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvtnt.bf16f32(<vscale x 8 x bfloat> %even, <vscale x 8 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvtnt_bf16, _f32, _x, )(even, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtnt_bf16_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvtnt.bf16f32(<vscale x 8 x bfloat> [[EVEN:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcvtnt_bf16_f32_mu14__SVBFloat16_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvtnt.bf16f32(<vscale x 8 x bfloat> [[EVEN:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svcvtnt_bf16_f32_m(svbfloat16_t even, svbool_t pg, svfloat32_t op) {
-  // CHECK-LABEL: test_svcvtnt_bf16_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.fcvtnt.bf16f32(<vscale x 8 x bfloat> %even, <vscale x 8 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svcvtnt_bf16, _f32, _m, )(even, pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_div.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_div.c
index b7e02df4c1520..4f384678c2431 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_div.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_div.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,415 +14,786 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdiv_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svdiv_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svdiv_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svdiv_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svdiv_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdiv_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdiv_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdiv_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdiv_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdiv_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdiv_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdiv_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdiv_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svdiv_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svdiv_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svdiv_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svdiv_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svdiv_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svdiv_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svdiv_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svdiv_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svdiv_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svdiv_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svdiv_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udiv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svdiv_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udiv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svdiv_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svdiv_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svdiv_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdiv_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdiv_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdiv_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdiv_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdiv_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdiv_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdiv_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svdiv_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svdiv_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svdiv_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svdiv_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svdiv_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svdiv_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svdiv_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svdiv_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdiv_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdiv_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svdiv_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svdiv_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdiv,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_divr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_divr.c
index 7bf735a49a046..b703082f0b0ef 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_divr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_divr.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,415 +14,786 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdivr_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svdivr_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svdivr_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svdivr_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svdivr_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdivr_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdivr_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdivr_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdivr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdivr_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdivr_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdivr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdivr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svdivr_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svdivr_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svdivr_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svdivr_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svdivr_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svdivr_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svdivr_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svdivr_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svdivr_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svdivr_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svdivr_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udivr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svdivr_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udivr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svdivr_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svdivr_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svdivr_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdivr_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdivr_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdivr_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdivr_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdivr_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdivr_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdivr_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svdivr_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svdivr_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svdivr_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svdivr_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svdivr_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svdivr_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svdivr_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fdivr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svdivr_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fdivr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svdivr_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdivr_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svdivr_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svdivr_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fdivr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdivr,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dot.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dot.c
index 8a3322df1b357..b08c6be1529a1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dot.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dot.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,118 +14,220 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdot_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svdot_s32u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svdot_s32(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svdot_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svdot_s64u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svdot_s64(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svdot_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svdot_u32u12__SVUint32_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svdot_u32(svuint32_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svdot_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svdot_u64u12__SVUint64_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svdot_u64(svuint64_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svdot_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdot_n_s32u11__SVInt32_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdot_n_s32(svint32_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svdot_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdot_n_s64u11__SVInt64_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdot_n_s64(svint64_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svdot_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdot_n_u32u12__SVUint32_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdot_n_u32(svuint32_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svdot_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdot_n_u64u12__SVUint64_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdot_n_u64(svuint64_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svdot_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot,_n_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svdot_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdot_lane_s32u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svdot_lane_s32(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svdot_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svdot_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svdot_lane_s32_1u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svdot_lane_s32_1(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svdot_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sdot.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot_lane,_s32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svdot_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdot_lane_s64u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svdot_lane_s64(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svdot_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svdot_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svdot_lane_s64_1u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svdot_lane_s64_1(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svdot_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sdot.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot_lane,_s64,,)(op1, op2, op3, 1);
 }
 
+// CHECK-LABEL: @test_svdot_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdot_lane_u32u12__SVUint32_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svdot_lane_u32(svuint32_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svdot_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.udot.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot_lane,_u32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svdot_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdot_lane_u64u12__SVUint64_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svdot_lane_u64(svuint64_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svdot_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.udot.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdot_lane,_u64,,)(op1, op2, op3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup-bfloat.c
index 65e6b93bdb4aa..b914fb6fc2cd9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,47 +15,86 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdup_n_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_bf16u6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svdup_n_bf16(bfloat16_t op) {
-  // CHECK-LABEL: test_svdup_n_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.x.nxv8bf16(bfloat %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svdup_n_bf16'}}
   return SVE_ACLE_FUNC(svdup, _n, _bf16, )(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_bf16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> [[TMP0]], bfloat [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_n_bf16_zu10__SVBool_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> [[TMP0]], bfloat [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svdup_n_bf16_z(svbool_t pg, bfloat16_t op) {
-  // CHECK-LABEL: test_svdup_n_bf16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> zeroinitializer, <vscale x 8 x i1> %[[PG]], bfloat %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svdup_n_bf16_z'}}
   return SVE_ACLE_FUNC(svdup, _n, _bf16_z, )(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_bf16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_n_bf16_mu14__SVBFloat16_tu10__SVBool_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svdup_n_bf16_m(svbfloat16_t inactive, svbool_t pg, bfloat16_t op) {
-  // CHECK-LABEL: test_svdup_n_bf16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> %inactive, <vscale x 8 x i1> %[[PG]], bfloat %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svdup_n_bf16_m'}}
   return SVE_ACLE_FUNC(svdup, _n, _bf16_m, )(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_bf16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> [[TMP0]], bfloat [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_n_bf16_xu10__SVBool_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> [[TMP0]], bfloat [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svdup_n_bf16_x(svbool_t pg, bfloat16_t op) {
-  // CHECK-LABEL: test_svdup_n_bf16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dup.nxv8bf16(<vscale x 8 x bfloat> undef, <vscale x 8 x i1> %[[PG]], bfloat %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svdup_n_bf16_x'}}
   return SVE_ACLE_FUNC(svdup, _n, _bf16_x, )(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_lane_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdup_lane_bf16u14__SVBFloat16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svdup_lane_bf16(svbfloat16_t data, uint16_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_bf16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl.nxv8bf16(<vscale x 8 x bfloat> %data, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svdup_lane_bf16'}}
   return SVE_ACLE_FUNC(svdup_lane,_bf16,,)(data, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup.c
index f632be7b85982..a09e952625701 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dup.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,515 +14,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdup_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svdup_n_s8a(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svdup_n_s8(int8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s8,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_s16s(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svdup_n_s16(int16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s16,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svdup_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s32,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svdup_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s64,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svdup_n_u8h(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svdup_n_u8(uint8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u8,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_u16t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svdup_n_u16(uint16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u16,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svdup_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u32,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svdup_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u64,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_f16Dh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svdup_n_f16(float16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f16,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_f32f(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svdup_n_f32(float32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f32,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_f64d(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svdup_n_f64(float64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f64,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_s8_zu10__SVBool_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svdup_n_s8_z(svbool_t pg, int8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s8_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s16_zu10__SVBool_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svdup_n_s16_z(svbool_t pg, int16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s16_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s32_zu10__SVBool_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdup_n_s32_z(svbool_t pg, int32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s32_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s64_zu10__SVBool_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdup_n_s64_z(svbool_t pg, int64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s64_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_u8_zu10__SVBool_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svdup_n_u8_z(svbool_t pg, uint8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u8_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u16_zu10__SVBool_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svdup_n_u16_z(svbool_t pg, uint16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u16_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u32_zu10__SVBool_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdup_n_u32_z(svbool_t pg, uint32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u32_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u64_zu10__SVBool_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdup_n_u64_z(svbool_t pg, uint64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u64_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], half [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f16_zu10__SVBool_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], half [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdup_n_f16_z(svbool_t pg, float16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], half %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f16_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], float [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f32_zu10__SVBool_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], float [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdup_n_f32_z(svbool_t pg, float32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], float %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f32_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], double [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f64_zu10__SVBool_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], double [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdup_n_f64_z(svbool_t pg, float64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], double %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f64_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_s8_mu10__SVInt8_tu10__SVBool_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svdup_n_s8_m(svint8_t inactive, svbool_t pg, int8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s8_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s16_mu11__SVInt16_tu10__SVBool_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svdup_n_s16_m(svint16_t inactive, svbool_t pg, int16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s16_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s32_mu11__SVInt32_tu10__SVBool_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdup_n_s32_m(svint32_t inactive, svbool_t pg, int32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s32_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s64_mu11__SVInt64_tu10__SVBool_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdup_n_s64_m(svint64_t inactive, svbool_t pg, int64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s64_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_u8_mu11__SVUint8_tu10__SVBool_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svdup_n_u8_m(svuint8_t inactive, svbool_t pg, uint8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u8_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u16_mu12__SVUint16_tu10__SVBool_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svdup_n_u16_m(svuint16_t inactive, svbool_t pg, uint16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u16_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u32_mu12__SVUint32_tu10__SVBool_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdup_n_u32_m(svuint32_t inactive, svbool_t pg, uint32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u32_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u64_mu12__SVUint64_tu10__SVBool_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdup_n_u64_m(svuint64_t inactive, svbool_t pg, uint64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u64_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], half [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f16_mu13__SVFloat16_tu10__SVBool_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], half [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdup_n_f16_m(svfloat16_t inactive, svbool_t pg, float16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], half %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f16_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], float [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f32_mu13__SVFloat32_tu10__SVBool_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], float [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdup_n_f32_m(svfloat32_t inactive, svbool_t pg, float32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], float %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f32_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], double [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f64_mu13__SVFloat64_tu10__SVBool_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], double [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdup_n_f64_m(svfloat64_t inactive, svbool_t pg, float64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], double %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f64_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_s8_xu10__SVBool_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svdup_n_s8_x(svbool_t pg, int8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s8_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s16_xu10__SVBool_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svdup_n_s16_x(svbool_t pg, int16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s16_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s32_xu10__SVBool_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdup_n_s32_x(svbool_t pg, int32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s32_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_s64_xu10__SVBool_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdup_n_s64_x(svbool_t pg, int64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_s64_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdup_n_u8_xu10__SVBool_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], i8 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svdup_n_u8_x(svbool_t pg, uint8_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, i8 %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u8_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u16_xu10__SVBool_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], i16 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svdup_n_u16_x(svbool_t pg, uint16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], i16 %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u16_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u32_xu10__SVBool_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], i32 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdup_n_u32_x(svbool_t pg, uint32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], i32 %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u32_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_u64_xu10__SVBool_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], i64 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdup_n_u64_x(svbool_t pg, uint64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], i64 %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_u64_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], half [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f16_xu10__SVBool_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], half [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdup_n_f16_x(svbool_t pg, float16_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], half %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f16_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], float [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f32_xu10__SVBool_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], float [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdup_n_f32_x(svbool_t pg, float32_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], float %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f32_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], double [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_n_f64_xu10__SVBool_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], double [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdup_n_f64_x(svbool_t pg, float64_t op)
 {
-  // CHECK-LABEL: test_svdup_n_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], double %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup,_n,_f64_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svdup_lane_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_lane_s8u10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svdup_lane_s8(svint8_t data, uint8_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_s8,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_s16u11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svdup_lane_s16(svint16_t data, uint16_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_s16,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_s32u11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svdup_lane_s32(svint32_t data, uint32_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_s32,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_s64u11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svdup_lane_s64(svint64_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_s64,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdup_lane_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svdup_lane_u8(svuint8_t data, uint8_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_u8,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svdup_lane_u16(svuint16_t data, uint16_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_u16,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svdup_lane_u32(svuint32_t data, uint32_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_u32,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svdup_lane_u64(svuint64_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_u64,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_f16u13__SVFloat16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svdup_lane_f16(svfloat16_t data, uint16_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_f16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl.nxv8f16(<vscale x 8 x half> %data, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_f16,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_f32u13__SVFloat32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svdup_lane_f32(svfloat32_t data, uint32_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_f32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_f32,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_lane_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[INDEX:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdup_lane_f64u13__SVFloat64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svdup_lane_f64(svfloat64_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdup_lane_f64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %index)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdup_lane,_f64,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdup_n_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.dup.x.nxv16i1(i1 [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svdup_n_b8b(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.dup.x.nxv16i1(i1 [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svdup_n_b8(bool op)
 {
-  // CHECK-LABEL: test_svdup_n_b8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.dup.x.nxv16i1(i1 %op)
-  // CHECK: ret <vscale x 16 x i1> %[[DUP]]
   return SVE_ACLE_FUNC(svdup,_n,_b8,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.dup.x.nxv8i1(i1 [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_b16b(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.dup.x.nxv8i1(i1 [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svdup_n_b16(bool op)
 {
-  // CHECK-LABEL: test_svdup_n_b16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.dup.x.nxv8i1(i1 %op)
-  // CHECK: %[[CVT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[CVT]]
   return SVE_ACLE_FUNC(svdup,_n,_b16,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.dup.x.nxv4i1(i1 [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_b32b(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.dup.x.nxv4i1(i1 [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svdup_n_b32(bool op)
 {
-  // CHECK-LABEL: test_svdup_n_b32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.dup.x.nxv4i1(i1 %op)
-  // CHECK: %[[CVT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[CVT]]
   return SVE_ACLE_FUNC(svdup,_n,_b32,)(op);
 }
 
+// CHECK-LABEL: @test_svdup_n_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.dup.x.nxv2i1(i1 [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdup_n_b64b(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.dup.x.nxv2i1(i1 [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svdup_n_b64(bool op)
 {
-  // CHECK-LABEL: test_svdup_n_b64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.dup.x.nxv2i1(i1 %op)
-  // CHECK: %[[CVT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i1> %[[CVT]]
   return SVE_ACLE_FUNC(svdup,_n,_b64,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
index 086d753870ec8..a1bc7df34c8ba 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,23 +15,51 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdupq_lane_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dupq.lane.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svdupq_lane_bf16u14__SVBFloat16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dupq.lane.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svdupq_lane_bf16(svbfloat16_t data, uint64_t index) {
-  // CHECK-LABEL: test_svdupq_lane_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dupq.lane.nxv8bf16(<vscale x 8 x bfloat> %data, i64 %index)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svdupq_lane_bf16'}}
   return SVE_ACLE_FUNC(svdupq_lane, _bf16, , )(data, index);
 }
+// CHECK-LABEL: @test_svdupq_n_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x bfloat> undef, bfloat [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x bfloat> [[TMP0]], bfloat [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x bfloat> [[TMP1]], bfloat [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x bfloat> [[TMP2]], bfloat [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x bfloat> [[TMP3]], bfloat [[X4:%.*]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x bfloat> [[TMP4]], bfloat [[X5:%.*]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x bfloat> [[TMP5]], bfloat [[X6:%.*]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x bfloat> [[TMP6]], bfloat [[X7:%.*]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x bfloat> @llvm.experimental.vector.insert.nxv8bf16.v8bf16(<vscale x 8 x bfloat> undef, <8 x bfloat> [[TMP7]], i64 0)
+// CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dupq.lane.nxv8bf16(<vscale x 8 x bfloat> [[TMP8]], i64 0)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP9]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svdupq_n_bf16u6__bf16u6__bf16u6__bf16u6__bf16u6__bf16u6__bf16u6__bf16u6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x bfloat> undef, bfloat [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x bfloat> [[TMP0]], bfloat [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x bfloat> [[TMP1]], bfloat [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x bfloat> [[TMP2]], bfloat [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x bfloat> [[TMP3]], bfloat [[X4:%.*]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x bfloat> [[TMP4]], bfloat [[X5:%.*]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x bfloat> [[TMP5]], bfloat [[X6:%.*]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x bfloat> [[TMP6]], bfloat [[X7:%.*]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x bfloat> @llvm.experimental.vector.insert.nxv8bf16.v8bf16(<vscale x 8 x bfloat> undef, <8 x bfloat> [[TMP7]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dupq.lane.nxv8bf16(<vscale x 8 x bfloat> [[TMP8]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP9]]
+//
 svbfloat16_t test_svdupq_n_bf16(bfloat16_t x0, bfloat16_t x1, bfloat16_t x2, bfloat16_t x3,
                                 bfloat16_t x4, bfloat16_t x5, bfloat16_t x6, bfloat16_t x7) {
-  // CHECK-LABEL: test_svdupq_n_bf16
-  // CHECK: insertelement <8 x bfloat> undef, bfloat %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <8 x bfloat> %[[X:.*]], bfloat %x7, i32 7
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 8 x bfloat> @llvm.experimental.vector.insert.nxv8bf16.v8bf16(<vscale x 8 x bfloat> undef, <8 x bfloat> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.dupq.lane.nxv8bf16(<vscale x 8 x bfloat> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 8 x bfloat> %[[DUPQ]]
   // expected-warning at +1 {{implicit declaration of function 'svdupq_n_bf16'}}
   return SVE_ACLE_FUNC(svdupq, _n, _bf16, )(x0, x1, x2, x3, x4, x5, x6, x7);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
index 680986c6ae0ba..effd3345303fe 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,317 +14,744 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svdupq_lane_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdupq_lane_s8u10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svdupq_lane_s8(svint8_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> %data, i64 %index)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_s8,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_s16u11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svdupq_lane_s16(svint16_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> %data, i64 %index)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_s16,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_s32u11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svdupq_lane_s32(svint32_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> %data, i64 %index)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_s32,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_s64u11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svdupq_lane_s64(svint64_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> %data, i64 %index)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_s64,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svdupq_lane_u8u11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svdupq_lane_u8(svuint8_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> %data, i64 %index)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_u8,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_u16u12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svdupq_lane_u16(svuint16_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> %data, i64 %index)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_u16,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_u32u12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svdupq_lane_u32(svuint32_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> %data, i64 %index)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_u32,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svdupq_lane_u64(svuint64_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> %data, i64 %index)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_u64,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_f16u13__SVFloat16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svdupq_lane_f16(svfloat16_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> %data, i64 %index)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_f16,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_f32u13__SVFloat32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svdupq_lane_f32(svfloat32_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> %data, i64 %index)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_f32,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_lane_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svdupq_lane_f64u13__SVFloat64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], i64 [[INDEX:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svdupq_lane_f64(svfloat64_t data, uint64_t index)
 {
-  // CHECK-LABEL: test_svdupq_lane_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> %data, i64 %index)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svdupq_lane,_f64,,)(data, index);
 }
 
+// CHECK-LABEL: @test_svdupq_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> undef, i8 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i32 8
+// CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i32 9
+// CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i32 10
+// CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i32 11
+// CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i32 12
+// CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i32 13
+// CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i32 14
+// CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i32 15
+// CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> [[TMP15]], i64 0)
+// CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdupq_n_s8aaaaaaaaaaaaaaaa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> undef, i8 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i32 8
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i32 9
+// CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i32 10
+// CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i32 11
+// CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i32 12
+// CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i32 13
+// CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i32 14
+// CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i32 15
+// CPP-CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> [[TMP15]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
+//
 svint8_t test_svdupq_n_s8(int8_t x0, int8_t x1, int8_t x2, int8_t x3,
                           int8_t x4, int8_t x5, int8_t x6, int8_t x7,
                           int8_t x8, int8_t x9, int8_t x10, int8_t x11,
                           int8_t x12, int8_t x13, int8_t x14, int8_t x15)
 {
-  // CHECK-LABEL: test_svdupq_n_s8
-  // CHECK: insertelement <16 x i8> undef, i8 %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <16 x i8> %[[X:.*]], i8 %x15, i32 15
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 16 x i8> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_s8,)(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15);
 }
 
+// CHECK-LABEL: @test_svdupq_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> undef, i16 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> [[TMP7]], i64 0)
+// CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_s16ssssssss(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> undef, i16 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> [[TMP7]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
+//
 svint16_t test_svdupq_n_s16(int16_t x0, int16_t x1, int16_t x2, int16_t x3,
                             int16_t x4, int16_t x5, int16_t x6, int16_t x7)
 {
-  // CHECK-LABEL: test_svdupq_n_s16
-  // CHECK: insertelement <8 x i16> undef, i16 %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <8 x i16> %[[X:.*]], i16 %x7, i32 7
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 8 x i16> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_s16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 }
 
+// CHECK-LABEL: @test_svdupq_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> undef, i32 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> [[TMP3]], i64 0)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_s32iiii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> undef, i32 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> [[TMP3]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
+//
 svint32_t test_svdupq_n_s32(int32_t x0, int32_t x1, int32_t x2, int32_t x3)
 {
-  // CHECK-LABEL: test_svdupq_n_s32
-  // CHECK: insertelement <4 x i32> undef, i32 %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <4 x i32> %[[X:.*]], i32 %x3, i32 3
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 4 x i32> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_s32,)(x0, x1, x2, x3);
 }
 
+// CHECK-LABEL: @test_svdupq_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> undef, i64 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> [[TMP1]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> undef, i64 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> [[TMP1]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svdupq_n_s64(int64_t x0, int64_t x1)
 {
-  // CHECK-LABEL: test_svdupq_n_s64
-  // CHECK: %[[SVEC:.*]] = insertelement <2 x i64> undef, i64 %x0, i32 0
-  // CHECK: %[[VEC:.*]] = insertelement <2 x i64> %[[SVEC]], i64 %x1, i32 1
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 2 x i64> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_s64,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svdupq_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> undef, i8 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i32 8
+// CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i32 9
+// CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i32 10
+// CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i32 11
+// CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i32 12
+// CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i32 13
+// CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i32 14
+// CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i32 15
+// CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> [[TMP15]], i64 0)
+// CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdupq_n_u8hhhhhhhhhhhhhhhh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> undef, i8 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[X4:%.*]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[X5:%.*]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[X6:%.*]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[X7:%.*]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[X8:%.*]], i32 8
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[X9:%.*]], i32 9
+// CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[X10:%.*]], i32 10
+// CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[X11:%.*]], i32 11
+// CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[X12:%.*]], i32 12
+// CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[X13:%.*]], i32 13
+// CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[X14:%.*]], i32 14
+// CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[X15:%.*]], i32 15
+// CPP-CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> [[TMP15]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP16]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP17]]
+//
 svuint8_t test_svdupq_n_u8(uint8_t x0, uint8_t x1, uint8_t x2, uint8_t x3,
                            uint8_t x4, uint8_t x5, uint8_t x6, uint8_t x7,
                            uint8_t x8, uint8_t x9, uint8_t x10, uint8_t x11,
                            uint8_t x12, uint8_t x13, uint8_t x14, uint8_t x15)
 {
-  // CHECK-LABEL: test_svdupq_n_u8
-  // CHECK: insertelement <16 x i8> undef, i8 %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <16 x i8> %[[X:.*]], i8 %x15, i32 15
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 16 x i8> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_u8,)(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15);
 }
 
+// CHECK-LABEL: @test_svdupq_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> undef, i16 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> [[TMP7]], i64 0)
+// CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_u16tttttttt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x i16> undef, i16 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x i16> [[TMP0]], i16 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x i16> [[TMP1]], i16 [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x i16> [[TMP3]], i16 [[X4:%.*]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x i16> [[TMP4]], i16 [[X5:%.*]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x i16> [[TMP5]], i16 [[X6:%.*]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x i16> [[TMP6]], i16 [[X7:%.*]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> [[TMP7]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP8]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP9]]
+//
 svuint16_t test_svdupq_n_u16(uint16_t x0, uint16_t x1, uint16_t x2, uint16_t x3,
                              uint16_t x4, uint16_t x5, uint16_t x6, uint16_t x7)
 {
-  // CHECK-LABEL: test_svdupq_n_u16
-  // CHECK: insertelement <8 x i16> undef, i16 %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <8 x i16> %[[X:.*]], i16 %x7, i32 7
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 8 x i16> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_u16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 }
 
+// CHECK-LABEL: @test_svdupq_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> undef, i32 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> [[TMP3]], i64 0)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_u32jjjj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> undef, i32 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> [[TMP3]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP4]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP5]]
+//
 svuint32_t test_svdupq_n_u32(uint32_t x0, uint32_t x1, uint32_t x2, uint32_t x3)
 {
-  // CHECK-LABEL: test_svdupq_n_u32
-  // CHECK: insertelement <4 x i32> undef, i32 %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <4 x i32> %[[X:.*]], i32 %x3, i32 3
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 4 x i32> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_u32,)(x0, x1, x2, x3);
 }
 
+// CHECK-LABEL: @test_svdupq_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> undef, i64 [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> [[TMP1]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> undef, i64 [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> [[TMP0]], i64 [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> [[TMP1]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP2]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svdupq_n_u64(uint64_t x0, uint64_t x1)
 {
-  // CHECK-LABEL: test_svdupq_n_u64
-  // CHECK: %[[SVEC:.*]] = insertelement <2 x i64> undef, i64 %x0, i32 0
-  // CHECK: %[[VEC:.*]] = insertelement <2 x i64> %[[SVEC]], i64 %x1, i32 1
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 2 x i64> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_u64,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svdupq_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x half> undef, half [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half [[X4:%.*]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half [[X5:%.*]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half [[X6:%.*]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half [[X7:%.*]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x half> @llvm.experimental.vector.insert.nxv8f16.v8f16(<vscale x 8 x half> undef, <8 x half> [[TMP7]], i64 0)
+// CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[TMP8]], i64 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP9]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_f16DhDhDhDhDhDhDhDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <8 x half> undef, half [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half [[X4:%.*]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half [[X5:%.*]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half [[X6:%.*]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half [[X7:%.*]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 8 x half> @llvm.experimental.vector.insert.nxv8f16.v8f16(<vscale x 8 x half> undef, <8 x half> [[TMP7]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> [[TMP8]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP9]]
+//
 svfloat16_t test_svdupq_n_f16(float16_t x0, float16_t x1, float16_t x2, float16_t x3,
                               float16_t x4, float16_t x5, float16_t x6, float16_t x7)
 {
-  // CHECK-LABEL: test_svdupq_n_f16
-  // CHECK: insertelement <8 x half> undef, half %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <8 x half> %[[X:.*]], half %x7, i32 7
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 8 x half> @llvm.experimental.vector.insert.nxv8f16.v8f16(<vscale x 8 x half> undef, <8 x half> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dupq.lane.nxv8f16(<vscale x 8 x half> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 8 x half> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_f16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 }
 
+// CHECK-LABEL: @test_svdupq_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x float> undef, float [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x float> [[TMP0]], float [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x float> [[TMP1]], float [[X2:%.*]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x float> [[TMP2]], float [[X3:%.*]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.experimental.vector.insert.nxv4f32.v4f32(<vscale x 4 x float> undef, <4 x float> [[TMP3]], i64 0)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[TMP4]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP5]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_f32ffff(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x float> undef, float [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x float> [[TMP0]], float [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x float> [[TMP1]], float [[X2:%.*]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x float> [[TMP2]], float [[X3:%.*]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.experimental.vector.insert.nxv4f32.v4f32(<vscale x 4 x float> undef, <4 x float> [[TMP3]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> [[TMP4]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP5]]
+//
 svfloat32_t test_svdupq_n_f32(float32_t x0, float32_t x1, float32_t x2, float32_t x3)
 {
-  // CHECK-LABEL: test_svdupq_n_f32
-  // CHECK: insertelement <4 x float> undef, float %x0, i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <4 x float> %[[X:.*]], float %x3, i32 3
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 4 x float> @llvm.experimental.vector.insert.nxv4f32.v4f32(<vscale x 4 x float> undef, <4 x float> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dupq.lane.nxv4f32(<vscale x 4 x float> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 4 x float> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_f32,)(x0, x1, x2, x3);
 }
 
+// CHECK-LABEL: @test_svdupq_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x double> undef, double [[X0:%.*]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x double> [[TMP0]], double [[X1:%.*]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v2f64(<vscale x 2 x double> undef, <2 x double> [[TMP1]], i64 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[TMP2]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_f64dd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x double> undef, double [[X0:%.*]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x double> [[TMP0]], double [[X1:%.*]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v2f64(<vscale x 2 x double> undef, <2 x double> [[TMP1]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> [[TMP2]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svdupq_n_f64(float64_t x0, float64_t x1)
 {
-  // CHECK-LABEL: test_svdupq_n_f64
-  // CHECK: %[[SVEC:.*]] = insertelement <2 x double> undef, double %x0, i32 0
-  // CHECK: %[[VEC:.*]] = insertelement <2 x double> %[[SVEC]], double %x1, i32 1
-  // CHECK-NOT: insertelement
-  // CHECK: %[[INS:.*]] = call <vscale x 2 x double> @llvm.experimental.vector.insert.nxv2f64.v2f64(<vscale x 2 x double> undef, <2 x double> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dupq.lane.nxv2f64(<vscale x 2 x double> %[[INS]], i64 0)
-  // CHECK: ret <vscale x 2 x double> %[[DUPQ]]
   return SVE_ACLE_FUNC(svdupq,_n,_f64,)(x0, x1);
 }
 
+// CHECK-LABEL: @test_svdupq_n_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL4:%.*]] = zext i1 [[X4:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL5:%.*]] = zext i1 [[X5:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL6:%.*]] = zext i1 [[X6:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL7:%.*]] = zext i1 [[X7:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL8:%.*]] = zext i1 [[X8:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL9:%.*]] = zext i1 [[X9:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL10:%.*]] = zext i1 [[X10:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL11:%.*]] = zext i1 [[X11:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL12:%.*]] = zext i1 [[X12:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL13:%.*]] = zext i1 [[X13:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL14:%.*]] = zext i1 [[X14:%.*]] to i8
+// CHECK-NEXT:    [[FROMBOOL15:%.*]] = zext i1 [[X15:%.*]] to i8
+// CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> undef, i8 [[FROMBOOL]], i32 0
+// CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[FROMBOOL1]], i32 1
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[FROMBOOL2]], i32 2
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[FROMBOOL3]], i32 3
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[FROMBOOL4]], i32 4
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[FROMBOOL5]], i32 5
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[FROMBOOL6]], i32 6
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[FROMBOOL7]], i32 7
+// CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[FROMBOOL8]], i32 8
+// CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[FROMBOOL9]], i32 9
+// CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[FROMBOOL10]], i32 10
+// CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[FROMBOOL11]], i32 11
+// CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[FROMBOOL12]], i32 12
+// CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[FROMBOOL13]], i32 13
+// CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[FROMBOOL14]], i32 14
+// CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[FROMBOOL15]], i32 15
+// CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> [[TMP15]], i64 0)
+// CHECK-NEXT:    [[TMP18:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP17]], i64 0)
+// CHECK-NEXT:    [[TMP19:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CHECK-NEXT:    [[TMP20:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[TMP16]], <vscale x 16 x i8> [[TMP18]], <vscale x 2 x i64> [[TMP19]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP20]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svdupq_n_b8bbbbbbbbbbbbbbbb(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[X0:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL1:%.*]] = zext i1 [[X1:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL2:%.*]] = zext i1 [[X2:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[X3:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL4:%.*]] = zext i1 [[X4:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL5:%.*]] = zext i1 [[X5:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL6:%.*]] = zext i1 [[X6:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL7:%.*]] = zext i1 [[X7:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL8:%.*]] = zext i1 [[X8:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL9:%.*]] = zext i1 [[X9:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL10:%.*]] = zext i1 [[X10:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL11:%.*]] = zext i1 [[X11:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL12:%.*]] = zext i1 [[X12:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL13:%.*]] = zext i1 [[X13:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL14:%.*]] = zext i1 [[X14:%.*]] to i8
+// CPP-CHECK-NEXT:    [[FROMBOOL15:%.*]] = zext i1 [[X15:%.*]] to i8
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = insertelement <16 x i8> undef, i8 [[FROMBOOL]], i32 0
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = insertelement <16 x i8> [[TMP0]], i8 [[FROMBOOL1]], i32 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <16 x i8> [[TMP1]], i8 [[FROMBOOL2]], i32 2
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <16 x i8> [[TMP2]], i8 [[FROMBOOL3]], i32 3
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <16 x i8> [[TMP3]], i8 [[FROMBOOL4]], i32 4
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <16 x i8> [[TMP4]], i8 [[FROMBOOL5]], i32 5
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <16 x i8> [[TMP5]], i8 [[FROMBOOL6]], i32 6
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <16 x i8> [[TMP6]], i8 [[FROMBOOL7]], i32 7
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <16 x i8> [[TMP7]], i8 [[FROMBOOL8]], i32 8
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <16 x i8> [[TMP8]], i8 [[FROMBOOL9]], i32 9
+// CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <16 x i8> [[TMP9]], i8 [[FROMBOOL10]], i32 10
+// CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <16 x i8> [[TMP10]], i8 [[FROMBOOL11]], i32 11
+// CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <16 x i8> [[TMP11]], i8 [[FROMBOOL12]], i32 12
+// CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <16 x i8> [[TMP12]], i8 [[FROMBOOL13]], i32 13
+// CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <16 x i8> [[TMP13]], i8 [[FROMBOOL14]], i32 14
+// CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <16 x i8> [[TMP14]], i8 [[FROMBOOL15]], i32 15
+// CPP-CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> [[TMP15]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP18:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> [[TMP17]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP19:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CPP-CHECK-NEXT:    [[TMP20:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> [[TMP16]], <vscale x 16 x i8> [[TMP18]], <vscale x 2 x i64> [[TMP19]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP20]]
+//
 svbool_t test_svdupq_n_b8(bool x0, bool x1, bool x2, bool x3,
                           bool x4, bool x5, bool x6, bool x7,
                           bool x8, bool x9, bool x10, bool x11,
                           bool x12, bool x13, bool x14, bool x15)
 {
-  // CHECK-LABEL: test_svdupq_n_b8
-  // CHECK-DAG: %[[X0:.*]] = zext i1 %x0 to i8
-  // CHECK-DAG: %[[X15:.*]] = zext i1 %x15 to i8
-  // CHECK: insertelement <16 x i8> undef, i8 %[[X0]], i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <16 x i8> %[[X:.*]], i8 %[[X15]], i32 15
-  // CHECK-NOT: insertelement
-  // CHECK: %[[PTRUE:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
-  // CHECK: %[[INS:.*]] = call <vscale x 16 x i8> @llvm.experimental.vector.insert.nxv16i8.v16i8(<vscale x 16 x i8> undef, <16 x i8> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dupq.lane.nxv16i8(<vscale x 16 x i8> %[[INS]], i64 0)
-  // CHECK: %[[ZERO:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
-  // CHECK: %[[CMP:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.cmpne.wide.nxv16i8(<vscale x 16 x i1> %[[PTRUE]], <vscale x 16 x i8> %[[DUPQ]], <vscale x 2 x i64> %[[ZERO]])
-  // CHECK: ret <vscale x 16 x i1> %[[CMP]]
   return SVE_ACLE_FUNC(svdupq,_n,_b8,)(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15);
 }
 
+// CHECK-LABEL: @test_svdupq_n_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[X0:%.*]] to i16
+// CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[X1:%.*]] to i16
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[X2:%.*]] to i16
+// CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[X3:%.*]] to i16
+// CHECK-NEXT:    [[TMP4:%.*]] = zext i1 [[X4:%.*]] to i16
+// CHECK-NEXT:    [[TMP5:%.*]] = zext i1 [[X5:%.*]] to i16
+// CHECK-NEXT:    [[TMP6:%.*]] = zext i1 [[X6:%.*]] to i16
+// CHECK-NEXT:    [[TMP7:%.*]] = zext i1 [[X7:%.*]] to i16
+// CHECK-NEXT:    [[TMP8:%.*]] = insertelement <8 x i16> undef, i16 [[TMP0]], i32 0
+// CHECK-NEXT:    [[TMP9:%.*]] = insertelement <8 x i16> [[TMP8]], i16 [[TMP1]], i32 1
+// CHECK-NEXT:    [[TMP10:%.*]] = insertelement <8 x i16> [[TMP9]], i16 [[TMP2]], i32 2
+// CHECK-NEXT:    [[TMP11:%.*]] = insertelement <8 x i16> [[TMP10]], i16 [[TMP3]], i32 3
+// CHECK-NEXT:    [[TMP12:%.*]] = insertelement <8 x i16> [[TMP11]], i16 [[TMP4]], i32 4
+// CHECK-NEXT:    [[TMP13:%.*]] = insertelement <8 x i16> [[TMP12]], i16 [[TMP5]], i32 5
+// CHECK-NEXT:    [[TMP14:%.*]] = insertelement <8 x i16> [[TMP13]], i16 [[TMP6]], i32 6
+// CHECK-NEXT:    [[TMP15:%.*]] = insertelement <8 x i16> [[TMP14]], i16 [[TMP7]], i32 7
+// CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
+// CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> [[TMP15]], i64 0)
+// CHECK-NEXT:    [[TMP18:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP17]], i64 0)
+// CHECK-NEXT:    [[TMP19:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CHECK-NEXT:    [[TMP20:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP16]], <vscale x 8 x i16> [[TMP18]], <vscale x 2 x i64> [[TMP19]])
+// CHECK-NEXT:    [[TMP21:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP20]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP21]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_b16bbbbbbbb(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[X0:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[X1:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[X2:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[X3:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext i1 [[X4:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = zext i1 [[X5:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = zext i1 [[X6:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = zext i1 [[X7:%.*]] to i16
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = insertelement <8 x i16> undef, i16 [[TMP0]], i32 0
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = insertelement <8 x i16> [[TMP8]], i16 [[TMP1]], i32 1
+// CPP-CHECK-NEXT:    [[TMP10:%.*]] = insertelement <8 x i16> [[TMP9]], i16 [[TMP2]], i32 2
+// CPP-CHECK-NEXT:    [[TMP11:%.*]] = insertelement <8 x i16> [[TMP10]], i16 [[TMP3]], i32 3
+// CPP-CHECK-NEXT:    [[TMP12:%.*]] = insertelement <8 x i16> [[TMP11]], i16 [[TMP4]], i32 4
+// CPP-CHECK-NEXT:    [[TMP13:%.*]] = insertelement <8 x i16> [[TMP12]], i16 [[TMP5]], i32 5
+// CPP-CHECK-NEXT:    [[TMP14:%.*]] = insertelement <8 x i16> [[TMP13]], i16 [[TMP6]], i32 6
+// CPP-CHECK-NEXT:    [[TMP15:%.*]] = insertelement <8 x i16> [[TMP14]], i16 [[TMP7]], i32 7
+// CPP-CHECK-NEXT:    [[TMP16:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP17:%.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> [[TMP15]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP18:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> [[TMP17]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP19:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CPP-CHECK-NEXT:    [[TMP20:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> [[TMP16]], <vscale x 8 x i16> [[TMP18]], <vscale x 2 x i64> [[TMP19]])
+// CPP-CHECK-NEXT:    [[TMP21:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP20]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP21]]
+//
 svbool_t test_svdupq_n_b16(bool x0, bool x1, bool x2, bool x3,
                            bool x4, bool x5, bool x6, bool x7)
 {
-  // CHECK-LABEL: test_svdupq_n_b16
-  // CHECK-DAG: %[[X0:.*]] = zext i1 %x0 to i16
-  // CHECK-DAG: %[[X7:.*]] = zext i1 %x7 to i16
-  // CHECK: insertelement <8 x i16> undef, i16 %[[X0]], i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <8 x i16> %[[X:.*]], i16 %[[X7]], i32 7
-  // CHECK-NOT: insertelement
-  // CHECK: %[[PTRUE:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
-  // CHECK: %[[INS:.*]] = call <vscale x 8 x i16> @llvm.experimental.vector.insert.nxv8i16.v8i16(<vscale x 8 x i16> undef, <8 x i16> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dupq.lane.nxv8i16(<vscale x 8 x i16> %[[INS]], i64 0)
-  // CHECK: %[[ZERO:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
-  // CHECK: %[[CMP:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.cmpne.wide.nxv8i16(<vscale x 8 x i1> %[[PTRUE]], <vscale x 8 x i16> %[[DUPQ]], <vscale x 2 x i64> %[[ZERO]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[CMP]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svdupq,_n,_b16,)(x0, x1, x2, x3, x4, x5, x6, x7);
 }
 
+// CHECK-LABEL: @test_svdupq_n_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[X0:%.*]] to i32
+// CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[X1:%.*]] to i32
+// CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[X2:%.*]] to i32
+// CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[X3:%.*]] to i32
+// CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> undef, i32 [[TMP0]], i32 0
+// CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x i32> [[TMP4]], i32 [[TMP1]], i32 1
+// CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> [[TMP5]], i32 [[TMP2]], i32 2
+// CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x i32> [[TMP6]], i32 [[TMP3]], i32 3
+// CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
+// CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> [[TMP7]], i64 0)
+// CHECK-NEXT:    [[TMP10:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP9]], i64 0)
+// CHECK-NEXT:    [[TMP11:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CHECK-NEXT:    [[TMP12:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP8]], <vscale x 4 x i32> [[TMP10]], <vscale x 2 x i64> [[TMP11]])
+// CHECK-NEXT:    [[TMP13:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP12]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP13]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_b32bbbb(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[X0:%.*]] to i32
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[X1:%.*]] to i32
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext i1 [[X2:%.*]] to i32
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext i1 [[X3:%.*]] to i32
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> undef, i32 [[TMP0]], i32 0
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x i32> [[TMP4]], i32 [[TMP1]], i32 1
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> [[TMP5]], i32 [[TMP2]], i32 2
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x i32> [[TMP6]], i32 [[TMP3]], i32 3
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> [[TMP7]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP10:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> [[TMP9]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP11:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CPP-CHECK-NEXT:    [[TMP12:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> [[TMP8]], <vscale x 4 x i32> [[TMP10]], <vscale x 2 x i64> [[TMP11]])
+// CPP-CHECK-NEXT:    [[TMP13:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP12]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP13]]
+//
 svbool_t test_svdupq_n_b32(bool x0, bool x1, bool x2, bool x3)
 {
-  // CHECK-LABEL: test_svdupq_n_b32
-  // CHECK-DAG: %[[X0:.*]] = zext i1 %x0 to i32
-  // CHECK-DAG: %[[X3:.*]] = zext i1 %x3 to i32
-  // CHECK: insertelement <4 x i32> undef, i32 %[[X0]], i32 0
   // <assume other insertelement>
-  // CHECK: %[[VEC:.*]] = insertelement <4 x i32> %[[X:.*]], i32 %[[X3]], i32 3
-  // CHECK-NOT: insertelement
-  // CHECK: %[[PTRUE:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
-  // CHECK: %[[INS:.*]] = call <vscale x 4 x i32> @llvm.experimental.vector.insert.nxv4i32.v4i32(<vscale x 4 x i32> undef, <4 x i32> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dupq.lane.nxv4i32(<vscale x 4 x i32> %[[INS]], i64 0)
-  // CHECK: %[[ZERO:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
-  // CHECK: %[[CMP:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.cmpne.wide.nxv4i32(<vscale x 4 x i1> %[[PTRUE]], <vscale x 4 x i32> %[[DUPQ]], <vscale x 2 x i64> %[[ZERO]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[CMP]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svdupq,_n,_b32,)(x0, x1, x2, x3);
 }
 
+// CHECK-LABEL: @test_svdupq_n_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[X0:%.*]] to i64
+// CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[X1:%.*]] to i64
+// CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i64> undef, i64 [[TMP0]], i32 0
+// CHECK-NEXT:    [[TMP3:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP1]], i32 1
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> [[TMP3]], i64 0)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP5]], i64 0)
+// CHECK-NEXT:    [[TMP7:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP4]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i64> [[TMP7]])
+// CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP8]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP9]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svdupq_n_b64bb(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = zext i1 [[X0:%.*]] to i64
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = zext i1 [[X1:%.*]] to i64
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i64> undef, i64 [[TMP0]], i32 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP1]], i32 1
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> [[TMP3]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> [[TMP5]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP7:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
+// CPP-CHECK-NEXT:    [[TMP8:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> [[TMP4]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i64> [[TMP7]])
+// CPP-CHECK-NEXT:    [[TMP9:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP8]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP9]]
+//
 svbool_t test_svdupq_n_b64(bool x0, bool x1)
 {
-  // CHECK-LABEL: test_svdupq_n_b64
-  // CHECK-DAG: %[[X0:.*]] = zext i1 %x0 to i64
-  // CHECK-DAG: %[[X1:.*]] = zext i1 %x1 to i64
-  // CHECK: %[[SVEC:.*]] = insertelement <2 x i64> undef, i64 %[[X0]], i32 0
-  // CHECK: %[[VEC:.*]] = insertelement <2 x i64> %[[SVEC]], i64 %[[X1]], i32 1
-  // CHECK-NOT: insertelement
-  // CHECK: %[[PTRUE:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
-  // CHECK: %[[INS:.*]] = call <vscale x 2 x i64> @llvm.experimental.vector.insert.nxv2i64.v2i64(<vscale x 2 x i64> undef, <2 x i64> %[[VEC]], i64 0)
-  // CHECK: %[[DUPQ:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dupq.lane.nxv2i64(<vscale x 2 x i64> %[[INS]], i64 0)
-  // CHECK: %[[ZERO:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 0)
-  // CHECK: %[[CMP:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.cmpne.nxv2i64(<vscale x 2 x i1> %[[PTRUE]], <vscale x 2 x i64> %[[DUPQ]], <vscale x 2 x i64> %[[ZERO]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[CMP]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svdupq,_n,_b64,)(x0, x1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eor.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eor.c
index 37c5a93ca4e04..66704fbdcec3c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eor.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eor.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,469 +14,889 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_sveor_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_sveor_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_sveor_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_sveor_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_sveor_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_sveor_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_sveor_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_sveor_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_sveor_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_sveor_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_sveor_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_sveor_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_sveor_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_sveor_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_sveor_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_sveor_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_sveor_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_sveor_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_sveor_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_sveor_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_sveor_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_sveor_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_sveor_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_sveor_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_sveor_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_sveor_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s8_z
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_sveor_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_sveor_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_sveor_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_sveor_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_sveor_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_sveor_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_sveor_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_sveor_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_sveor_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_sveor_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_sveor_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_sveor_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_sveor_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_sveor_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_sveor_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_sveor_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_sveor_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_sveor_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_sveor_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_sveor_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_sveor_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_sveor_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveor_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_sveor_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_sveor_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_sveor_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.eor.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_sveor_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.eor.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_sveor_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_sveor_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.eor.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveor,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eorv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eorv.c
index a6fedfc48f5b8..c2d9f658d75eb 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eorv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_eorv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,72 +14,134 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_sveorv_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.eorv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_sveorv_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.eorv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_sveorv_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_sveorv_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.eorv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.eorv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorv_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.eorv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_sveorv_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_sveorv_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.eorv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.eorv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorv_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.eorv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_sveorv_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_sveorv_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.eorv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.eorv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorv_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.eorv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_sveorv_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_sveorv_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.eorv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.eorv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_sveorv_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.eorv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_sveorv_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_sveorv_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.eorv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.eorv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorv_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.eorv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_sveorv_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_sveorv_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.eorv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.eorv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorv_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.eorv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_sveorv_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_sveorv_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.eorv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_sveorv_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.eorv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorv_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.eorv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_sveorv_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_sveorv_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.eorv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(sveorv,_u64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_expa.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_expa.c
index 946c8feb51322..bb43fe46cc7d0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_expa.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_expa.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svexpa_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fexpa.x.nxv8f16(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svexpa_f16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fexpa.x.nxv8f16(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svexpa_f16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svexpa_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fexpa.x.nxv8f16(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexpa,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svexpa_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fexpa.x.nxv4f32(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svexpa_f32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fexpa.x.nxv4f32(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svexpa_f32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svexpa_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fexpa.x.nxv4f32(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexpa,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svexpa_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fexpa.x.nxv2f64(<vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svexpa_f64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fexpa.x.nxv2f64(<vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svexpa_f64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svexpa_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fexpa.x.nxv2f64(<vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexpa,_f64,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext-bfloat.c
index 94efa175f8f42..7612697ce9a0b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext-bfloat.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1  -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1  -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svext_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ext.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i32 127)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svext_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ext.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i32 127)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svext_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svext_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ext.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2, i32 127)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svext_bf16'}}
   return SVE_ACLE_FUNC(svext,_bf16,,)(op1, op2, 127);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext.c
index b4bfd4c875158..170ab7f53bc3a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ext.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,122 +14,227 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svext_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svext_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svext_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svext_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s8,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svext_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 255)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svext_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 255)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svext_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svext_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 255)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s8,,)(op1, op2, 255);
 }
 
+// CHECK-LABEL: @test_svext_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svext_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svext_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svext_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 127)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svext_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 127)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svext_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svext_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 127)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s16,,)(op1, op2, 127);
 }
 
+// CHECK-LABEL: @test_svext_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svext_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svext_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svext_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svext_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svext_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svext_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 63)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s32,,)(op1, op2, 63);
 }
 
+// CHECK-LABEL: @test_svext_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svext_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svext_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svext_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svext_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svext_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svext_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 31)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_s64,,)(op1, op2, 31);
 }
 
+// CHECK-LABEL: @test_svext_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 255)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svext_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 255)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svext_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svext_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ext.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 255)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_u8,,)(op1, op2, 255);
 }
 
+// CHECK-LABEL: @test_svext_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 127)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 127)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svext_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svext_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ext.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 127)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_u16,,)(op1, op2, 127);
 }
 
+// CHECK-LABEL: @test_svext_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svext_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svext_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ext.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 63)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_u32,,)(op1, op2, 63);
 }
 
+// CHECK-LABEL: @test_svext_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svext_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svext_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ext.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 31)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_u64,,)(op1, op2, 31);
 }
 
+// CHECK-LABEL: @test_svext_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ext.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 127)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ext.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 127)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svext_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svext_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ext.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 127)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_f16,,)(op1, op2, 127);
 }
 
+// CHECK-LABEL: @test_svext_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ext.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ext.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svext_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svext_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ext.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, i32 63)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_f32,,)(op1, op2, 63);
 }
 
+// CHECK-LABEL: @test_svext_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ext.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svext_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ext.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svext_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svext_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ext.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, i32 31)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svext,_f64,,)(op1, op2, 31);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extb.c
index b9dea71280fdd..890469e8afe04 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,164 +14,308 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svextb_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svextb_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svextb_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svextb_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svextb_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svextb_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svextb_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svextb_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svextb_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svextb_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svextb_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svextb_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svextb_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svextb_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svextb_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svextb_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svextb_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svextb_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svextb_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svextb_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svextb_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svextb_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svextb_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svextb_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svextb_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svextb_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svextb_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sxtb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svextb_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svextb_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxtb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svextb_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svextb_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svextb_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svextb_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uxtb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svextb_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svextb_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxtb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextb_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextb_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svextb_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svextb_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextb,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_exth.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_exth.c
index 09a2de0c29da2..00bb66679c0a3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_exth.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_exth.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,110 +14,206 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svexth_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svexth_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svexth_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svexth_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svexth_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svexth_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svexth_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svexth_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svexth_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svexth_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svexth_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svexth_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svexth_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svexth_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svexth_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svexth_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svexth_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svexth_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svexth_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sxth.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svexth_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svexth_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxth.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svexth_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svexth_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uxth.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svexth_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svexth_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svexth_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svexth_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxth.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svexth,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extw.c
index 4dd352568131d..cebc35e36dae1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_extw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,56 +14,104 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svextw_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextw_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svextw_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svextw_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextw,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextw_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextw_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svextw_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svextw_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextw,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextw_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextw_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svextw_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svextw_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextw,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextw_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextw_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svextw_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svextw_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextw,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svextw_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextw_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svextw_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svextw_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sxtw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextw,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svextw_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svextw_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svextw_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svextw_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uxtw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svextw,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
index 0c8a8ad19ecc0..2a2327ddf6b4b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,20 +14,34 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svget2_bf16_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget2_bf16_014svbfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget2_bf16_0(svbfloat16x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_bf16_0
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget2_bf16'}}
   return SVE_ACLE_FUNC(svget2,_bf16,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget2_bf16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget2_bf16_114svbfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget2_bf16_1(svbfloat16x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_bf16_1
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget2_bf16'}}
   return SVE_ACLE_FUNC(svget2,_bf16,,)(tuple, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
index 55e85bc06e3bb..d9d013e133906 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get2.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,90 +14,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svget2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svget2_s810svint8x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svget2_s8(svint8x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_s8
-  // CHECK: %[[EXT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_s8,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_s1611svint16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svget2_s16(svint16x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_s16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_s16,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_s3211svint32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svget2_s32(svint32x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_s32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_s32,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_s6411svint64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svget2_s64(svint64x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_s64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_s64,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svget2_u811svuint8x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svget2_u8(svuint8x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_u8
-  // CHECK: %[[EXT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_u8,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_u1612svuint16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svget2_u16(svuint16x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_u16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_u16,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_u3212svuint32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svget2_u32(svuint32x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_u32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_u32,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_u6412svuint64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svget2_u64(svuint64x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_u64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_u64,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_f1613svfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svget2_f16(svfloat16x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_f16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 8 x half> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_f16,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_f3213svfloat32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svget2_f32(svfloat32x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_f32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 4 x float> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_f32,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget2_f6413svfloat64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svget2_f64(svfloat64x2_t tuple)
 {
-  // CHECK-LABEL: test_svget2_f64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 2 x double> %[[EXT]]
   return SVE_ACLE_FUNC(svget2,_f64,,)(tuple, 0);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
index c4efd621a3367..cffcce28e4226 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,29 +14,50 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svget3_bf16_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget3_bf16_014svbfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget3_bf16_0(svbfloat16x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_bf16_0
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget3_bf16'}}
   return SVE_ACLE_FUNC(svget3,_bf16,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget3_bf16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget3_bf16_114svbfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget3_bf16_1(svbfloat16x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_bf16_1
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget3_bf16'}}
   return SVE_ACLE_FUNC(svget3,_bf16,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget3_bf16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget3_bf16_214svbfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget3_bf16_2(svbfloat16x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_bf16_2
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget3_bf16'}}
   return SVE_ACLE_FUNC(svget3,_bf16,,)(tuple, 2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
index e4c2315c2cc2a..a1f26d9732769 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get3.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,90 +14,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svget3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svget3_s810svint8x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svget3_s8(svint8x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_s8
-  // CHECK: %[[EXT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_s8,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_s1611svint16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svget3_s16(svint16x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_s16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_s16,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_s3211svint32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svget3_s32(svint32x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_s32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_s32,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_s6411svint64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svget3_s64(svint64x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_s64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_s64,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svget3_u811svuint8x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svget3_u8(svuint8x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_u8
-  // CHECK: %[[EXT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_u8,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_u1612svuint16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svget3_u16(svuint16x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_u16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_u16,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_u3212svuint32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svget3_u32(svuint32x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_u32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_u32,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_u6412svuint64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svget3_u64(svuint64x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_u64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_u64,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget3_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_f1613svfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svget3_f16(svfloat16x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_f16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x half> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_f16,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_f3213svfloat32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svget3_f32(svfloat32x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_f32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 4 x float> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_f32,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget3_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget3_f6413svfloat64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svget3_f64(svfloat64x3_t tuple)
 {
-  // CHECK-LABEL: test_svget3_f64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 2 x double> %[[EXT]]
   return SVE_ACLE_FUNC(svget3,_f64,,)(tuple, 2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
index b14583058382a..95b4600d37183 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,38 +14,66 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svget4_bf16_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget4_bf16_014svbfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget4_bf16_0(svbfloat16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_bf16_0
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget4_bf16'}}
   return SVE_ACLE_FUNC(svget4,_bf16,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget4_bf16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget4_bf16_114svbfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget4_bf16_1(svbfloat16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_bf16_1
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %tuple, i32 1)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget4_bf16'}}
   return SVE_ACLE_FUNC(svget4,_bf16,,)(tuple, 1);
 }
 
+// CHECK-LABEL: @test_svget4_bf16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget4_bf16_214svbfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget4_bf16_2(svbfloat16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_bf16_2
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget4_bf16'}}
   return SVE_ACLE_FUNC(svget4,_bf16,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget4_bf16_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svget4_bf16_314svbfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svget4_bf16_3(svbfloat16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_bf16_3
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %tuple, i32 3)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[EXT]]
   // expected-warning at +1 {{implicit declaration of function 'svget4_bf16'}}
   return SVE_ACLE_FUNC(svget4,_bf16,,)(tuple, 3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
index 00094d7465688..76c39645a67e0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_get4.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -15,90 +16,167 @@
 
 // NOTE: For these tests clang converts the struct parameter into
 // several parameters, one for each member of the original struct.
+// CHECK-LABEL: @test_svget4_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svget4_s810svint8x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svget4_s8(svint8x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_s8
-  // CHECK: %[[EXT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_s8,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget4_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_s1611svint16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svget4_s16(svint16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_s16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_s16,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget4_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_s3211svint32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svget4_s32(svint32x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_s32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_s32,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget4_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_s6411svint64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svget4_s64(svint64x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_s64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %tuple, i32 3)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_s64,,)(tuple, 3);
 }
 
+// CHECK-LABEL: @test_svget4_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svget4_u811svuint8x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svget4_u8(svuint8x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_u8
-  // CHECK: %[[EXT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_u8,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget4_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_u1612svuint16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svget4_u16(svuint16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_u16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %tuple, i32 3)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_u16,,)(tuple, 3);
 }
 
+// CHECK-LABEL: @test_svget4_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_u3212svuint32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svget4_u32(svuint32x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_u32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_u32,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget4_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_u6412svuint64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svget4_u64(svuint64x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_u64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %tuple, i32 3)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_u64,,)(tuple, 3);
 }
 
+// CHECK-LABEL: @test_svget4_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_f1613svfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svget4_f16(svfloat16x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_f16
-  // CHECK: %[[EXT:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 8 x half> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_f16,,)(tuple, 2);
 }
 
+// CHECK-LABEL: @test_svget4_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_f3213svfloat32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svget4_f32(svfloat32x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_f32
-  // CHECK: %[[EXT:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %tuple, i32 0)
-  // CHECK-NEXT: ret <vscale x 4 x float> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_f32,,)(tuple, 0);
 }
 
+// CHECK-LABEL: @test_svget4_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svget4_f6413svfloat64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svget4_f64(svfloat64x4_t tuple)
 {
-  // CHECK-LABEL: test_svget4_f64
-  // CHECK: %[[EXT:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %tuple, i32 2)
-  // CHECK-NEXT: ret <vscale x 2 x double> %[[EXT]]
   return SVE_ACLE_FUNC(svget4,_f64,,)(tuple, 2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_index.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_index.c
index 96cbada5663a5..23cfd38d0eb40 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_index.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_index.c
@@ -1,69 +1,126 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svindex_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.index.nxv16i8(i8 [[BASE:%.*]], i8 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svindex_s8aa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.index.nxv16i8(i8 [[BASE:%.*]], i8 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svindex_s8(int8_t base, int8_t step)
 {
-  // CHECK-LABEL: test_svindex_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.index.nxv16i8(i8 %base, i8 %step)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return svindex_s8(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.index.nxv8i16(i16 [[BASE:%.*]], i16 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svindex_s16ss(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.index.nxv8i16(i16 [[BASE:%.*]], i16 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svindex_s16(int16_t base, int16_t step)
 {
-  // CHECK-LABEL: test_svindex_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.index.nxv8i16(i16 %base, i16 %step)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return svindex_s16(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.index.nxv4i32(i32 [[BASE:%.*]], i32 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svindex_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.index.nxv4i32(i32 [[BASE:%.*]], i32 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svindex_s32(int32_t base, int32_t step)
 {
-  // CHECK-LABEL: test_svindex_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.index.nxv4i32(i32 %base, i32 %step)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return svindex_s32(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.index.nxv2i64(i64 [[BASE:%.*]], i64 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svindex_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.index.nxv2i64(i64 [[BASE:%.*]], i64 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svindex_s64(int64_t base, int64_t step)
 {
-  // CHECK-LABEL: test_svindex_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.index.nxv2i64(i64 %base, i64 %step)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return svindex_s64(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.index.nxv16i8(i8 [[BASE:%.*]], i8 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svindex_u8hh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.index.nxv16i8(i8 [[BASE:%.*]], i8 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svindex_u8(uint8_t base, uint8_t step)
 {
-  // CHECK-LABEL: test_svindex_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.index.nxv16i8(i8 %base, i8 %step)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return svindex_u8(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.index.nxv8i16(i16 [[BASE:%.*]], i16 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svindex_u16tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.index.nxv8i16(i16 [[BASE:%.*]], i16 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svindex_u16(uint16_t base, uint16_t step)
 {
-  // CHECK-LABEL: test_svindex_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.index.nxv8i16(i16 %base, i16 %step)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return svindex_u16(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.index.nxv4i32(i32 [[BASE:%.*]], i32 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svindex_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.index.nxv4i32(i32 [[BASE:%.*]], i32 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svindex_u32(uint32_t base, uint32_t step)
 {
-  // CHECK-LABEL: test_svindex_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.index.nxv4i32(i32 %base, i32 %step)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return svindex_u32(base, step);
 }
 
+// CHECK-LABEL: @test_svindex_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.index.nxv2i64(i64 [[BASE:%.*]], i64 [[STEP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svindex_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.index.nxv2i64(i64 [[BASE:%.*]], i64 [[STEP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svindex_u64(uint64_t base, uint64_t step)
 {
-  // CHECK-LABEL: test_svindex_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.index.nxv2i64(i64 %base, i64 %step)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return svindex_u64(base, step);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr-bfloat.c
index 330d0c3fe00fd..7385dc51e0472 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,10 +15,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svinsr_n_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.insr.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], bfloat [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svinsr_n_bf16u14__SVBFloat16_tu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.insr.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], bfloat [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svinsr_n_bf16(svbfloat16_t op1, bfloat16_t op2) {
-  // CHECK-LABEL: test_svinsr_n_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.insr.nxv8bf16(<vscale x 8 x bfloat> %op1, bfloat %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svinsr_n_bf16'}}
   return SVE_ACLE_FUNC(svinsr, _n_bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr.c
index b1d6c86a4d5a9..20a4e43d8ad06 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_insr.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,90 +14,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svinsr_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.insr.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], i8 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svinsr_n_s8u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.insr.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svinsr_n_s8(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.insr.nxv16i8(<vscale x 16 x i8> %op1, i8 %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.insr.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i16 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.insr.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svinsr_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.insr.nxv8i16(<vscale x 8 x i16> %op1, i16 %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.insr.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.insr.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svinsr_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.insr.nxv4i32(<vscale x 4 x i32> %op1, i32 %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.insr.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.insr.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svinsr_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.insr.nxv2i64(<vscale x 2 x i64> %op1, i64 %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.insr.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], i8 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svinsr_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.insr.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svinsr_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.insr.nxv16i8(<vscale x 16 x i8> %op1, i8 %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.insr.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i16 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.insr.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svinsr_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.insr.nxv8i16(<vscale x 8 x i16> %op1, i16 %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.insr.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.insr.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svinsr_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.insr.nxv4i32(<vscale x 4 x i32> %op1, i32 %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.insr.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.insr.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svinsr_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.insr.nxv2i64(<vscale x 2 x i64> %op1, i64 %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.insr.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], half [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_f16u13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.insr.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svinsr_n_f16(svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.insr.nxv8f16(<vscale x 8 x half> %op1, half %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.insr.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], float [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_f32u13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.insr.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svinsr_n_f32(svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.insr.nxv4f32(<vscale x 4 x float> %op1, float %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svinsr_n_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.insr.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], double [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svinsr_n_f64u13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.insr.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svinsr_n_f64(svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svinsr_n_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.insr.nxv2f64(<vscale x 2 x double> %op1, double %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svinsr,_n_f64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta-bfloat.c
index 8a9a9b3011421..029d7b648ff01 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,11 +15,19 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlasta_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.lasta.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CHECK-NEXT:    ret bfloat [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlasta_bf16u10__SVBool_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.lasta.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret bfloat [[TMP1]]
+//
 bfloat16_t test_svlasta_bf16(svbool_t pg, svbfloat16_t op) {
-  // CHECK-LABEL: test_svlasta_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call bfloat @llvm.aarch64.sve.lasta.nxv8bf16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op)
-  // CHECK: ret bfloat %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svlasta_bf16'}}
   return SVE_ACLE_FUNC(svlasta, _bf16, , )(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta.c
index 08c599d849689..e22f5f7808d1e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lasta.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,99 +14,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlasta_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlasta_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svlasta_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svlasta_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.lasta.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svlasta_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svlasta_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.lasta.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svlasta_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svlasta_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.lasta.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svlasta_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svlasta_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.lasta.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlasta_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lasta.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svlasta_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svlasta_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.lasta.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lasta.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svlasta_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svlasta_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.lasta.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lasta.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svlasta_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svlasta_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.lasta.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lasta.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlasta_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svlasta_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.lasta.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_u64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.lasta.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.lasta.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svlasta_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svlasta_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.lasta.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.lasta.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.lasta.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svlasta_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svlasta_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.lasta.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlasta_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.lasta.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlasta_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.lasta.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svlasta_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svlasta_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.lasta.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlasta,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb-bfloat.c
index ac5ca3db00c09..de02399167089 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,11 +15,19 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlastb_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.lastb.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CHECK-NEXT:    ret bfloat [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlastb_bf16u10__SVBool_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call bfloat @llvm.aarch64.sve.lastb.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret bfloat [[TMP1]]
+//
 bfloat16_t test_svlastb_bf16(svbool_t pg, svbfloat16_t op) {
-  // CHECK-LABEL: test_svlastb_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call bfloat @llvm.aarch64.sve.lastb.nxv8bf16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op)
-  // CHECK: ret bfloat %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svlastb_bf16'}}
   return SVE_ACLE_FUNC(svlastb, _bf16, , )(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb.c
index 2d29af299317a..7d3ce15dbe901 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lastb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,99 +14,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlastb_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlastb_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svlastb_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svlastb_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.lastb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svlastb_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svlastb_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.lastb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svlastb_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svlastb_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.lastb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svlastb_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svlastb_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.lastb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlastb_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.lastb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svlastb_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svlastb_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.lastb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.lastb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svlastb_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svlastb_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.lastb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.lastb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svlastb_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svlastb_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.lastb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.lastb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlastb_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svlastb_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.lastb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_u64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.lastb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.lastb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svlastb_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svlastb_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.lastb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.lastb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.lastb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svlastb_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svlastb_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.lastb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlastb_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.lastb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlastb_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.lastb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svlastb_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svlastb_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.lastb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlastb,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c
index 55d9761c46050..88c39af449635 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,24 +14,42 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld1_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svld1_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svld1_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svld1_bf16'}}
   return SVE_ACLE_FUNC(svld1,_bf16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svld1_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
 svbfloat16_t test_svld1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svld1_vnum_bf16'}}
   return SVE_ACLE_FUNC(svld1_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1.c
index 4e7225c10244d..f3833761b41a1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,560 +14,1096 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld1_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svld1_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svld1_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld1_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld1_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svld1_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld1_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svld1_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svld1_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld1_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld1_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svld1_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svld1_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svld1_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svld1_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld1_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svld1_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld1_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svld1_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_s8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svld1_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svld1_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svld1_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld1_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svld1_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_u8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svld1_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svld1_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svld1_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svld1_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svld1_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld1_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svld1_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1_vnum_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_vnum,_f64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svld1_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1_gather_u32base_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svld1_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svld1_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1_gather_u32base_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svld1_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svld1_gather_u32base_f32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_gather_u32base_f32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1_gather_u32base_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _f32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z29test_svld1_gather_u64base_f64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_gather_u64base_f64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1_gather_u64base_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _f64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_s32offset_s32u10__SVBool_tPKiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_gather_s32offset_s32(svbool_t pg, const int32_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_s32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s32, offset, _s32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_s64offset_s64u10__SVBool_tPKlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_gather_s64offset_s64(svbool_t pg, const int64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s64, offset, _s64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_s32offset_u32u10__SVBool_tPKju11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_gather_s32offset_u32(svbool_t pg, const uint32_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_s32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s32, offset, _u32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_s64offset_u64u10__SVBool_tPKmu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_gather_s64offset_u64(svbool_t pg, const uint64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s64, offset, _u64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_s32offset_f32u10__SVBool_tPKfu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_gather_s32offset_f32(svbool_t pg, const float32_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_s32offset_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s32, offset, _f32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_s64offset_f64u10__SVBool_tPKdu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_gather_s64offset_f64(svbool_t pg, const float64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_s64offset_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s64, offset, _f64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_u32offset_s32u10__SVBool_tPKiu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_gather_u32offset_s32(svbool_t pg, const int32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_u32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u32, offset, _s32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_u64offset_s64u10__SVBool_tPKlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_gather_u64offset_s64(svbool_t pg, const int64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u64, offset, _s64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_u32offset_u32u10__SVBool_tPKju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_gather_u32offset_u32(svbool_t pg, const uint32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_u32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u32, offset, _u32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_u64offset_u64u10__SVBool_tPKmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_gather_u64offset_u64(svbool_t pg, const uint64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u64, offset, _u64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_u32offset_f32u10__SVBool_tPKfu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_gather_u32offset_f32(svbool_t pg, const float32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_u32offset_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u32, offset, _f32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1_gather_u64offset_f64u10__SVBool_tPKdu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_gather_u64offset_f64(svbool_t pg, const float64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1_gather_u64offset_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u64, offset, _f64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z36test_svld1_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1_gather_u32base_offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z36test_svld1_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z36test_svld1_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1_gather_u32base_offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z36test_svld1_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z36test_svld1_gather_u32base_offset_f32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_gather_u32base_offset_f32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1_gather_u32base_offset_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _offset_f32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z36test_svld1_gather_u64base_offset_f64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_gather_u64base_offset_f64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1_gather_u64base_offset_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _offset_f64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_s32index_s32u10__SVBool_tPKiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_gather_s32index_s32(svbool_t pg, const int32_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1_gather_s32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s32, index, _s32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_s64index_s64u10__SVBool_tPKlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_gather_s64index_s64(svbool_t pg, const int64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1_gather_s64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s64, index, _s64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_s32index_u32u10__SVBool_tPKju11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_gather_s32index_u32(svbool_t pg, const uint32_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1_gather_s32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s32, index, _u32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_s64index_u64u10__SVBool_tPKmu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_gather_s64index_u64(svbool_t pg, const uint64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1_gather_s64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s64, index, _u64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s32index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_s32index_f32u10__SVBool_tPKfu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_gather_s32index_f32(svbool_t pg, const float32_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1_gather_s32index_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s32, index, _f32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_s64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_s64index_f64u10__SVBool_tPKdu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_gather_s64index_f64(svbool_t pg, const float64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1_gather_s64index_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.index.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, s64, index, _f64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_u32index_s32u10__SVBool_tPKiu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1_gather_u32index_s32(svbool_t pg, const int32_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1_gather_u32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u32, index, _s32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_u64index_s64u10__SVBool_tPKlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1_gather_u64index_s64(svbool_t pg, const int64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1_gather_u64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u64, index, _s64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_u32index_u32u10__SVBool_tPKju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1_gather_u32index_u32(svbool_t pg, const uint32_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1_gather_u32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u32, index, _u32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_u64index_u64u10__SVBool_tPKmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1_gather_u64index_u64(svbool_t pg, const uint64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1_gather_u64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u64, index, _u64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_u32index_f32u10__SVBool_tPKfu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1_gather_u32index_f32(svbool_t pg, const float32_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1_gather_u32index_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u32, index, _f32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z30test_svld1_gather_u64index_f64u10__SVBool_tPKdu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1_gather_u64index_f64(svbool_t pg, const float64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1_gather_u64index_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.index.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather_, u64, index, _f64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svld1_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1_gather_u32base_index_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svld1_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1_gather_u64base_index_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svld1_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1_gather_u32base_index_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svld1_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1_gather_u64base_index_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _index_u64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u32base_index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svld1_gather_u32base_index_f32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svld1_gather_u32base_index_f32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1_gather_u32base_index_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u32base, _index_f32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1_gather_u64base_index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svld1_gather_u64base_index_f64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svld1_gather_u64base_index_f64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1_gather_u64base_index_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld1_gather, _u64base, _index_f64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro-bfloat.c
index 254797e28c9ac..d15fdb5ab4dc2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,10 +13,18 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1ro_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1ro.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svld1ro_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1ro.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svld1ro_bf16(svbool_t pg, const bfloat16_t *base) {
-  // CHECK-LABEL: test_svld1ro_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1ro.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _bf16, , )(pg, base);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro.c
index 8f2e7aab912b8..222aeb3f31c76 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ro.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +f64mm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,88 +13,174 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1ro_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1ro.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld1ro_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1ro.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svld1ro_s8(svbool_t pg, const int8_t *base) {
-  // CHECK-LABEL: test_svld1ro_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1ro.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _s8, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1ro.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1ro.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svld1ro_s16(svbool_t pg, const int16_t *base) {
-  // CHECK-LABEL: test_svld1ro_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1ro.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _s16, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1ro.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1ro.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1ro_s32(svbool_t pg, const int32_t *base) {
-  // CHECK-LABEL: test_svld1ro_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1ro.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _s32, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1ro.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1ro.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1ro_s64(svbool_t pg, const int64_t *base) {
-  // CHECK-LABEL: test_svld1ro_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1ro.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _s64, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1ro.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld1ro_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1ro.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svld1ro_u8(svbool_t pg, const uint8_t *base) {
-  // CHECK-LABEL: test_svld1ro_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1ro.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _u8, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1ro.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1ro.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svld1ro_u16(svbool_t pg, const uint16_t *base) {
-  // CHECK-LABEL: test_svld1ro_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1ro.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _u16, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1ro.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1ro.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1ro_u32(svbool_t pg, const uint32_t *base) {
-  // CHECK-LABEL: test_svld1ro_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1ro.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _u32, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1ro.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1ro.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1ro_u64(svbool_t pg, const uint64_t *base) {
-  // CHECK-LABEL: test_svld1ro_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1ro.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _u64, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1ro.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1ro.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svld1ro_f16(svbool_t pg, const float16_t *base) {
-  // CHECK-LABEL: test_svld1ro_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1ro.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _f16, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1ro.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1ro.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1ro_f32(svbool_t pg, const float32_t *base) {
-  // CHECK-LABEL: test_svld1ro_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1ro.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _f32, , )(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ro_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1ro.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ro_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1ro.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1ro_f64(svbool_t pg, const float64_t *base) {
-  // CHECK-LABEL: test_svld1ro_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1ro.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1ro, _f64, , )(pg, base);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c
index bf1688bb01180..da35a474bf8ec 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,12 +14,20 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1rq_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svld1rq_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svld1rq_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svld1rq_bf16'}}
   return SVE_ACLE_FUNC(svld1rq,_bf16,,)(pg, base);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
index f3bc8467d1bab..9f8cbf82c9bb8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,99 +15,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1rq_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld1rq_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svld1rq_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svld1rq_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svld1rq_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svld1rq_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld1rq_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svld1rq_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svld1rq_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svld1rq_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svld1rq_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1rq.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1rq.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svld1rq_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1rq.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1rq.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1rq.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svld1rq_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1rq.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1rq_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1rq.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1rq_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1rq.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svld1rq_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svld1rq_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1rq.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svld1rq,_f64,,)(pg, base);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sb.c
index d2c72d684aa3f..b3b04310192f7 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,278 +14,542 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1sb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sb_s16u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svld1sb_s16(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1sb_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svld1sb_s16(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sb_s32u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sb_s32(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1sb_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sb_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sb_s64u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sb_s64(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1sb_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sb_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sb_u16u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svld1sb_u16(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1sb_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svld1sb_u16(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sb_u32u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sb_u32(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1sb_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sb_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sb_u64u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sb_u64(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld1sb_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sb_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sb_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sb_vnum_s16u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svint16_t test_svld1sb_vnum_s16(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sb_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svld1sb_vnum_s16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sb_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sb_vnum_s32u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svld1sb_vnum_s32(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sb_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sb_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sb_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sb_vnum_s64u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svld1sb_vnum_s64(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sb_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sb_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sb_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sb_vnum_u16u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svuint16_t test_svld1sb_vnum_u16(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sb_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svld1sb_vnum_u16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sb_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sb_vnum_u32u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svld1sb_vnum_u32(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sb_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sb_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sb_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sb_vnum_u64u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svld1sb_vnum_u64(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sb_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sb_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sb_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sb_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1sb_gather_u32base_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sb_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sb_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1sb_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sb_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sb_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1sb_gather_u32base_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sb_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sb_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1sb_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_s32offset_s32u10__SVBool_tPKau11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sb_gather_s32offset_s32(svbool_t pg, const int8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_s32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_s64offset_s64u10__SVBool_tPKau11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sb_gather_s64offset_s64(svbool_t pg, const int8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_s32offset_u32u10__SVBool_tPKau11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sb_gather_s32offset_u32(svbool_t pg, const int8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_s32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_s64offset_u64u10__SVBool_tPKau11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sb_gather_s64offset_u64(svbool_t pg, const int8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_u32offset_s32u10__SVBool_tPKau12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sb_gather_u32offset_s32(svbool_t pg, const int8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_u32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_u64offset_s64u10__SVBool_tPKau12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sb_gather_u64offset_s64(svbool_t pg, const int8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_u32offset_u32u10__SVBool_tPKau12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sb_gather_u32offset_u32(svbool_t pg, const int8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_u32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sb_gather_u64offset_u64u10__SVBool_tPKau12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sb_gather_u64offset_u64(svbool_t pg, const int8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1sb_gather_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sb_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sb_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sb_gather_u32base_offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sb_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sb_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sb_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sb_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sb_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sb_gather_u32base_offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sb_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sb_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sb_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sb_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sb_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sh.c
index acf716b06a503..cc6b1d35d053d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,346 +14,682 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1sh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sh_s32u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_s32(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld1sh_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sh_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sh_s64u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_s64(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld1sh_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sh_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sh_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sh_u32u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_u32(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld1sh_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sh_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sh_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sh_u64u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_u64(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld1sh_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sh_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sh_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sh_vnum_s32u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svld1sh_vnum_s32(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sh_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sh_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sh_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sh_vnum_s64u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svld1sh_vnum_s64(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sh_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sh_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sh_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sh_vnum_u32u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svld1sh_vnum_u32(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sh_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svld1sh_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sh_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sh_vnum_u64u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svld1sh_vnum_u64(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sh_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sh_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sh_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1sh_gather_u32base_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sh_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1sh_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sh_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1sh_gather_u32base_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sh_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1sh_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_s32offset_s32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_gather_s32offset_s32(svbool_t pg, const int16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_s32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_s64offset_s64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_gather_s64offset_s64(svbool_t pg, const int16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_s32offset_u32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_gather_s32offset_u32(svbool_t pg, const int16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_s32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_s64offset_u64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_gather_s64offset_u64(svbool_t pg, const int16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_u32offset_s32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_gather_u32offset_s32(svbool_t pg, const int16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_u32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_u64offset_s64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_gather_u64offset_s64(svbool_t pg, const int16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_u64offset_s64
-  // CHECK: %[[PG]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_u32offset_u32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_gather_u32offset_u32(svbool_t pg, const int16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_u32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sh_gather_u64offset_u64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_gather_u64offset_u64(svbool_t pg, const int16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1sh_gather_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sh_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sh_gather_u32base_offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sh_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sh_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sh_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sh_gather_u32base_offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %2
   return SVE_ACLE_FUNC(svld1sh_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sh_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sh_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_s32index_s32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_gather_s32index_s32(svbool_t pg, const int16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_s32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_s64index_s64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_gather_s64index_s64(svbool_t pg, const int16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_s64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_s32index_u32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_gather_s32index_u32(svbool_t pg, const int16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_s32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_s64index_u64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_gather_s64index_u64(svbool_t pg, const int16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_s64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_u32index_s32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1sh_gather_u32index_s32(svbool_t pg, const int16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_u32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_u64index_s64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sh_gather_u64index_s64(svbool_t pg, const int16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_u64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_u32index_u32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1sh_gather_u32index_u32(svbool_t pg, const int16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_u32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sh_gather_u64index_u64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sh_gather_u64index_u64(svbool_t pg, const int16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1sh_gather_u64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]]  = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]]  = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1sh_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svld1sh_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1sh_gather_u32base_index_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1sh_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svld1sh_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1sh_gather_u64base_index_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1sh_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svld1sh_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1sh_gather_u32base_index_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1sh_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1sh_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svld1sh_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1sh_gather_u64base_index_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sh_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sw.c
index b9c4dcdaff600..ce6daaffd286a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1sw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,174 +14,342 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1sw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sw_s64u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_s64(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld1sw_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sw_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1sw_u64u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_u64(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld1sw_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sw_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1sw_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sw_vnum_s64u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svld1sw_vnum_s64(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sw_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sw_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sw_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1sw_vnum_u64u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svld1sw_vnum_u64(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1sw_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svld1sw_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sw_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1sw_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1sw_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1sw_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sw_gather_s64offset_s64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_gather_s64offset_s64(svbool_t pg, const int32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1sw_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sw_gather_s64offset_u64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_gather_s64offset_u64(svbool_t pg, const int32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1sw_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sw_gather_u64offset_s64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_gather_u64offset_s64(svbool_t pg, const int32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1sw_gather_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1sw_gather_u64offset_u64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_gather_u64offset_u64(svbool_t pg, const int32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1sw_gather_u64offset_u64
-  // CHECK: %[[PG]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sw_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sw_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1sw_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1sw_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sw_gather_s64index_s64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_gather_s64index_s64(svbool_t pg, const int32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1sw_gather_s64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sw_gather_s64index_u64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_gather_s64index_u64(svbool_t pg, const int32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1sw_gather_s64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sw_gather_u64index_s64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1sw_gather_u64index_s64(svbool_t pg, const int32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1sw_gather_u64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]]  = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]]  = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1sw_gather_u64index_u64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1sw_gather_u64index_u64(svbool_t pg, const int32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1sw_gather_u64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1sw_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svld1sw_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1sw_gather_u64base_index_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1sw_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1sw_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svld1sw_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1sw_gather_u64base_index_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: <vscale x 2 x i64> %[[SEXT]]
   return SVE_ACLE_FUNC(svld1sw_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ub.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ub.c
index 6df00731b60af..73f95358a59dc 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ub.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1ub.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,278 +14,542 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1ub_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ub_s16u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svld1ub_s16(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1ub_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svld1ub_s16(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ub_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ub_s32u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1ub_s32(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1ub_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1ub_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ub_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ub_s64u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1ub_s64(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1ub_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1ub_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ub_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ub_u16u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svld1ub_u16(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1ub_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svld1ub_u16(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ub_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ub_u32u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1ub_u32(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1ub_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1ub_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ub_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1ub_u64u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1ub_u64(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld1ub_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1ub_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1ub_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1ub_vnum_s16u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svint16_t test_svld1ub_vnum_s16(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1ub_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svld1ub_vnum_s16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1ub_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1ub_vnum_s32u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svld1ub_vnum_s32(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1ub_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1ub_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1ub_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1ub_vnum_s64u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svld1ub_vnum_s64(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1ub_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1ub_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1ub_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1ub_vnum_u16u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svuint16_t test_svld1ub_vnum_u16(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1ub_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svld1ub_vnum_u16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1ub_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1ub_vnum_u32u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svld1ub_vnum_u32(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1ub_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1ub_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1ub_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1ub_vnum_u64u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svld1ub_vnum_u64(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1ub_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1ub_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1ub_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1ub_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1ub_gather_u32base_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1ub_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1ub_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1ub_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1ub_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1ub_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1ub_gather_u32base_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1ub_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1ub_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1ub_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_s32offset_s32u10__SVBool_tPKhu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1ub_gather_s32offset_s32(svbool_t pg, const uint8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_s32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_s64offset_s64u10__SVBool_tPKhu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1ub_gather_s64offset_s64(svbool_t pg, const uint8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_s32offset_u32u10__SVBool_tPKhu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1ub_gather_s32offset_u32(svbool_t pg, const uint8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_s32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_s64offset_u64u10__SVBool_tPKhu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1ub_gather_s64offset_u64(svbool_t pg, const uint8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_u32offset_s32u10__SVBool_tPKhu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1ub_gather_u32offset_s32(svbool_t pg, const uint8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_u32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_u64offset_s64u10__SVBool_tPKhu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1ub_gather_u64offset_s64(svbool_t pg, const uint8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_u32offset_u32u10__SVBool_tPKhu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1ub_gather_u32offset_u32(svbool_t pg, const uint8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_u32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1ub_gather_u64offset_u64u10__SVBool_tPKhu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1ub_gather_u64offset_u64(svbool_t pg, const uint8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1ub_gather_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1ub_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1ub_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1ub_gather_u32base_offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1ub_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1ub_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1ub_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1ub_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1ub_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1ub_gather_u32base_offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1ub_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1ub_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1ub_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1ub_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1ub_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uh.c
index 735868396ae06..7b9a2af0f7331 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,346 +14,682 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1uh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1uh_s32u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_s32(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld1uh_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1uh_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1uh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1uh_s64u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_s64(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld1uh_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uh_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1uh_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1uh_u32u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_u32(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld1uh_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1uh_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1uh_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1uh_u64u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_u64(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld1uh_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uh_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1uh_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1uh_vnum_s32u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svld1uh_vnum_s32(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1uh_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1uh_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1uh_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1uh_vnum_s64u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svld1uh_vnum_s64(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1uh_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uh_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1uh_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1uh_vnum_u32u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svld1uh_vnum_u32(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1uh_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svld1uh_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1uh_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1uh_vnum_u64u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svld1uh_vnum_u64(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1uh_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uh_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1uh_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1uh_gather_u32base_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1uh_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1uh_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1uh_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svld1uh_gather_u32base_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: et <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1uh_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1uh_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_s32offset_s32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_gather_s32offset_s32(svbool_t pg, const uint16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_s32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_s64offset_s64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_gather_s64offset_s64(svbool_t pg, const uint16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %1 to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_s32offset_u32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_gather_s32offset_u32(svbool_t pg, const uint16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_s32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_s64offset_u64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_gather_s64offset_u64(svbool_t pg, const uint16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %1 to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_u32offset_s32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_gather_u32offset_s32(svbool_t pg, const uint16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_u32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_u64offset_s64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_gather_u64offset_s64(svbool_t pg, const uint16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_u32offset_u32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_gather_u32offset_u32(svbool_t pg, const uint16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_u32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uh_gather_u64offset_u64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_gather_u64offset_u64(svbool_t pg, const uint16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1uh_gather_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %1 to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1uh_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1uh_gather_u32base_offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1uh_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1uh_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1uh_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1uh_gather_u32base_offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %2
   return SVE_ACLE_FUNC(svld1uh_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1uh_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1uh_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_s32index_s32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_gather_s32index_s32(svbool_t pg, const uint16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_s32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_s64index_s64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_gather_s64index_s64(svbool_t pg, const uint16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_s64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_s32index_u32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_gather_s32index_u32(svbool_t pg, const uint16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_s32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_s64index_u64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_gather_s64index_u64(svbool_t pg, const uint16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_s64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_u32index_s32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svld1uh_gather_u32index_s32(svbool_t pg, const uint16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_u32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_u64index_s64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uh_gather_u64index_s64(svbool_t pg, const uint16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_u64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_u32index_u32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svld1uh_gather_u32index_u32(svbool_t pg, const uint16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_u32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uh_gather_u64index_u64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uh_gather_u64index_u64(svbool_t pg, const uint16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1uh_gather_u64index_u64
-  // CHECK: %[[PG]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.index.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1uh_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svld1uh_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1uh_gather_u32base_index_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1uh_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svld1uh_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1uh_gather_u64base_index_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1uh_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svld1uh_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1uh_gather_u32base_index_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1uh_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1uh_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svld1uh_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1uh_gather_u64base_index_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uh_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uw.c
index e487d08a5c902..fb7f1f89266a9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1uw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,174 +14,342 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld1uw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1uw_s64u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_s64(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld1uw_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uw_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1uw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svld1uw_u64u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_u64(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld1uw_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uw_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svld1uw_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1uw_vnum_s64u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svld1uw_vnum_s64(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1uw_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uw_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1uw_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svld1uw_vnum_u64u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svld1uw_vnum_u64(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld1uw_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svld1uw_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1uw_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1uw_gather_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svld1uw_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svld1uw_gather_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uw_gather_s64offset_s64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_gather_s64offset_s64(svbool_t pg, const uint32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1uw_gather_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uw_gather_s64offset_u64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_gather_s64offset_u64(svbool_t pg, const uint32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svld1uw_gather_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uw_gather_u64offset_s64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_gather_u64offset_s64(svbool_t pg, const uint32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1uw_gather_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svld1uw_gather_u64offset_u64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_gather_u64offset_u64(svbool_t pg, const uint32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svld1uw_gather_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1uw_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1uw_gather_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svld1uw_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svld1uw_gather_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uw_gather_s64index_s64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_gather_s64index_s64(svbool_t pg, const uint32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1uw_gather_s64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %1 to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uw_gather_s64index_u64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_gather_s64index_u64(svbool_t pg, const uint32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svld1uw_gather_s64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %1 to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uw_gather_u64index_s64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svld1uw_gather_u64index_s64(svbool_t pg, const uint32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1uw_gather_u64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svld1uw_gather_u64index_u64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svld1uw_gather_u64index_u64(svbool_t pg, const uint32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svld1uw_gather_u64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.index.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1uw_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svld1uw_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1uw_gather_u64base_index_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svld1uw_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svld1uw_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svld1uw_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svld1uw_gather_u64base_index_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ld1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: <vscale x 2 x i64> %[[ZEXT]]
   return SVE_ACLE_FUNC(svld1uw_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2-bfloat.c
index 4fa4a0692f5c1..f3adc9c882af2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,23 +13,41 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.ld2.nxv16bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld2_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.ld2.nxv16bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP1]]
+//
 svbfloat16x2_t test_svld2_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svld2_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.ld2.nxv16bf16.nxv8i1(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret <vscale x 16 x bfloat> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_bf16,,)(pg, base);
 }
 
 
+// CHECK-LABEL: @test_svld2_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.ld2.nxv16bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svld2_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.ld2.nxv16bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP3]]
+//
 svbfloat16x2_t test_svld2_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.ld2.nxv16bf16.nxv8i1(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x bfloat> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2.c
index fbb994b4a891d..303ad2bcb0a2f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld2.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,218 +13,412 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld2_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
 svint8x2_t test_svld2_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld2_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
+//
 svint16x2_t test_svld2_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld2_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
+//
 svint32x2_t test_svld2_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld2_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
+//
 svint64x2_t test_svld2_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svld2_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld2_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
 svuint8x2_t test_svld2_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld2_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP1]]
+//
 svuint16x2_t test_svld2_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld2_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP1]]
+//
 svuint32x2_t test_svld2_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld2_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP1]]
+//
 svuint64x2_t test_svld2_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svld2_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.ld2.nxv16f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.ld2.nxv16f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP1]]
+//
 svfloat16x2_t test_svld2_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svld2_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.ld2.nxv16f16.nxv8i1(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret <vscale x 16 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.ld2.nxv8f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.ld2.nxv8f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP1]]
+//
 svfloat32x2_t test_svld2_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svld2_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.ld2.nxv8f32.nxv4i1(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret <vscale x 8 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.ld2.nxv4f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld2_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.ld2.nxv4f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP1]]
+//
 svfloat64x2_t test_svld2_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svld2_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.ld2.nxv4f64.nxv2i1(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret <vscale x 4 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld2_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP2]]
+//
 svint8x2_t test_svld2_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_s8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP3]]
+//
 svint16x2_t test_svld2_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP3]]
+//
 svint32x2_t test_svld2_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP3]]
+//
 svint64x2_t test_svld2_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld2_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP2]]
+//
 svuint8x2_t test_svld2_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_u8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.ld2.nxv32i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP3]]
+//
 svuint16x2_t test_svld2_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.ld2.nxv16i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP3]]
+//
 svuint32x2_t test_svld2_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.ld2.nxv8i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP3]]
+//
 svuint64x2_t test_svld2_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.ld2.nxv4i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.ld2.nxv16f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.ld2.nxv16f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP3]]
+//
 svfloat16x2_t test_svld2_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.ld2.nxv16f16.nxv8i1(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.ld2.nxv8f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.ld2.nxv8f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP3]]
+//
 svfloat32x2_t test_svld2_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.ld2.nxv8f32.nxv4i1(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 8 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld2_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.ld2.nxv4f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld2_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.ld2.nxv4f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP3]]
+//
 svfloat64x2_t test_svld2_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld2_vnum_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.ld2.nxv4f64.nxv2i1(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 4 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld2_vnum,_f64,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3-bfloat.c
index a5afd64089bed..318c6d82d1a90 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,22 +13,40 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld3_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.ld3.nxv24bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld3_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.ld3.nxv24bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP1]]
+//
 svbfloat16x3_t test_svld3_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svld3_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.ld3.nxv24bf16.nxv8i1(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret <vscale x 24 x bfloat> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_bf16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.ld3.nxv24bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svld3_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.ld3.nxv24bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP3]]
+//
 svbfloat16x3_t test_svld3_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.ld3.nxv24bf16.nxv8i1(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 24 x bfloat> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3.c
index 9353dae26f0d0..e207bece5574c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld3.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,218 +13,412 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld3_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
 svint8x3_t test_svld3_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld3_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP1]]
+//
 svint16x3_t test_svld3_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld3_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP1]]
+//
 svint32x3_t test_svld3_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld3_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP1]]
+//
 svint64x3_t test_svld3_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svld3_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld3_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
 svuint8x3_t test_svld3_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld3_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP1]]
+//
 svuint16x3_t test_svld3_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld3_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP1]]
+//
 svuint32x3_t test_svld3_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld3_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP1]]
+//
 svuint64x3_t test_svld3_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svld3_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.ld3.nxv24f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.ld3.nxv24f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP1]]
+//
 svfloat16x3_t test_svld3_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svld3_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.ld3.nxv24f16.nxv8i1(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret <vscale x 24 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.ld3.nxv12f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.ld3.nxv12f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP1]]
+//
 svfloat32x3_t test_svld3_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svld3_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.ld3.nxv12f32.nxv4i1(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret <vscale x 12 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.ld3.nxv6f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld3_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.ld3.nxv6f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP1]]
+//
 svfloat64x3_t test_svld3_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svld3_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.ld3.nxv6f64.nxv2i1(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret <vscale x 6 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld3_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
+//
 svint8x3_t test_svld3_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_s8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP3]]
+//
 svint16x3_t test_svld3_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP3]]
+//
 svint32x3_t test_svld3_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP3]]
+//
 svint64x3_t test_svld3_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld3_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP2]]
+//
 svuint8x3_t test_svld3_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_u8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.ld3.nxv48i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP3]]
+//
 svuint16x3_t test_svld3_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.ld3.nxv24i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP3]]
+//
 svuint32x3_t test_svld3_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.ld3.nxv12i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP3]]
+//
 svuint64x3_t test_svld3_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.ld3.nxv6i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.ld3.nxv24f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 24 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.ld3.nxv24f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP3]]
+//
 svfloat16x3_t test_svld3_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.ld3.nxv24f16.nxv8i1(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 24 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.ld3.nxv12f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 12 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.ld3.nxv12f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP3]]
+//
 svfloat32x3_t test_svld3_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.ld3.nxv12f32.nxv4i1(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 12 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld3_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.ld3.nxv6f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 6 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld3_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.ld3.nxv6f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP3]]
+//
 svfloat64x3_t test_svld3_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld3_vnum_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.ld3.nxv6f64.nxv2i1(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 6 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld3_vnum,_f64,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4-bfloat.c
index 9d73d5f88f3e8..483a7b7b5dc1c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,22 +13,40 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld4_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.ld4.nxv32bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svld4_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.ld4.nxv32bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP1]]
+//
 svbfloat16x4_t test_svld4_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svld4_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.ld4.nxv32bf16.nxv8i1(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_bf16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.ld4.nxv32bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svld4_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.ld4.nxv32bf16.nxv8i1(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP3]]
+//
 svbfloat16x4_t test_svld4_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.ld4.nxv32bf16.nxv8i1(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4.c
index bff9ae56ed601..bee97c9474410 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld4.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,218 +13,412 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svld4_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld4_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
 svint8x4_t test_svld4_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svld4_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP1]]
+//
 svint16x4_t test_svld4_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svld4_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP1]]
+//
 svint32x4_t test_svld4_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svld4_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP1]]
+//
 svint64x4_t test_svld4_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svld4_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svld4_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
 svuint8x4_t test_svld4_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svld4_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP1]]
+//
 svuint16x4_t test_svld4_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svld4_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP1]]
+//
 svuint32x4_t test_svld4_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svld4_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP1]]
+//
 svuint64x4_t test_svld4_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svld4_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.ld4.nxv32f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.ld4.nxv32f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP1]]
+//
 svfloat16x4_t test_svld4_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svld4_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.ld4.nxv32f16.nxv8i1(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret <vscale x 32 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.ld4.nxv16f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.ld4.nxv16f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP1]]
+//
 svfloat32x4_t test_svld4_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svld4_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.ld4.nxv16f32.nxv4i1(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret <vscale x 16 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.ld4.nxv8f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svld4_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.ld4.nxv8f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP1]]
+//
 svfloat64x4_t test_svld4_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svld4_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.ld4.nxv8f64.nxv2i1(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret <vscale x 8 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld4_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP2]]
+//
 svint8x4_t test_svld4_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_s8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
+//
 svint16x4_t test_svld4_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
+//
 svint32x4_t test_svld4_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
+//
 svint64x4_t test_svld4_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svld4_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP2]]
+//
 svuint8x4_t test_svld4_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_u8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.ld4.nxv64i8.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP3]]
+//
 svuint16x4_t test_svld4_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.ld4.nxv32i16.nxv8i1(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP3]]
+//
 svuint32x4_t test_svld4_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.ld4.nxv16i32.nxv4i1(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP3]]
+//
 svuint64x4_t test_svld4_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.ld4.nxv8i64.nxv2i1(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.ld4.nxv32f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 32 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.ld4.nxv32f16.nxv8i1(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP3]]
+//
 svfloat16x4_t test_svld4_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.ld4.nxv32f16.nxv8i1(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 32 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.ld4.nxv16f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.ld4.nxv16f32.nxv4i1(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP3]]
+//
 svfloat32x4_t test_svld4_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.ld4.nxv16f32.nxv4i1(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 16 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svld4_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.ld4.nxv8f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svld4_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.ld4.nxv8f64.nxv2i1(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP3]]
+//
 svfloat64x4_t test_svld4_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svld4_vnum_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.ld4.nxv8f64.nxv2i1(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret <vscale x 8 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svld4_vnum,_f64,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c
index 848a7a8fcbb35..89b0992cd53da 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,24 +14,42 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svldff1_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svldff1_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svldff1_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svldff1_bf16'}}
   return SVE_ACLE_FUNC(svldff1,_bf16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svldff1_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
 svbfloat16_t test_svldff1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_bf16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svldff1_vnum_bf16'}}
   return SVE_ACLE_FUNC(svldff1_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1.c
index 0ded34cc7cb82..ca12ca0610af0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,560 +14,1096 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svldff1_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svldff1_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svldff1_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldff1_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldff1_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svldff1_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svldff1_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svldff1_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svldff1_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldff1_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldff1_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svldff1_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svldff1_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svldff1_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svldff1_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldff1_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svldff1_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svldff1_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svldff1_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_s8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svldff1_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_s16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldff1_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_s32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldff1_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_s64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svldff1_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svldff1_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_u8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldff1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svldff1_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_u16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldff1_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_u32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldff1_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_u64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svldff1_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_f16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svldff1_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_f32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldff1_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svldff1_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1_vnum_f64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svldff1_vnum,_f64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldff1_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldff1_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldff1_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldff1_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldff1_gather_u32base_f32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_gather_u32base_f32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _f32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldff1_gather_u64base_f64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_gather_u64base_f64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _f64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_s32offset_s32u10__SVBool_tPKiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_gather_s32offset_s32(svbool_t pg, const int32_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_s32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s32, offset, _s32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_s64offset_s64u10__SVBool_tPKlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_gather_s64offset_s64(svbool_t pg, const int64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s64, offset, _s64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_s32offset_u32u10__SVBool_tPKju11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_gather_s32offset_u32(svbool_t pg, const uint32_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_s32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s32, offset, _u32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_s64offset_u64u10__SVBool_tPKmu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_gather_s64offset_u64(svbool_t pg, const uint64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s64, offset, _u64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_s32offset_f32u10__SVBool_tPKfu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_gather_s32offset_f32(svbool_t pg, const float32_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_s32offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4f32(<vscale x 4 x i1> [[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s32, offset, _f32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_s64offset_f64u10__SVBool_tPKdu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_gather_s64offset_f64(svbool_t pg, const float64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_s64offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s64, offset, _f64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_u32offset_s32u10__SVBool_tPKiu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_gather_u32offset_s32(svbool_t pg, const int32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u32, offset, _s32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_u64offset_s64u10__SVBool_tPKlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_gather_u64offset_s64(svbool_t pg, const int64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u64, offset, _s64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_u32offset_u32u10__SVBool_tPKju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_gather_u32offset_u32(svbool_t pg, const uint32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u32, offset, _u32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_u64offset_u64u10__SVBool_tPKmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_gather_u64offset_u64(svbool_t pg, const uint64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u64, offset, _u64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_u32offset_f32u10__SVBool_tPKfu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_gather_u32offset_f32(svbool_t pg, const float32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_u32offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u32, offset, _f32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1_gather_u64offset_f64u10__SVBool_tPKdu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_gather_u64offset_f64(svbool_t pg, const float64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1_gather_u64offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u64, offset, _f64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldff1_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldff1_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldff1_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldff1_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldff1_gather_u32base_offset_f32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_gather_u32base_offset_f32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _offset_f32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldff1_gather_u64base_offset_f64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_gather_u64base_offset_f64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _offset_f64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_s32index_s32u10__SVBool_tPKiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_gather_s32index_s32(svbool_t pg, const int32_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_s32index_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s32, index, _s32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_s64index_s64u10__SVBool_tPKlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_gather_s64index_s64(svbool_t pg, const int64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s64, index, _s64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_s32index_u32u10__SVBool_tPKju11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_gather_s32index_u32(svbool_t pg, const uint32_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_s32index_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s32, index, _u32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_s64index_u64u10__SVBool_tPKmu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_gather_s64index_u64(svbool_t pg, const uint64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s64, index, _u64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s32index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_s32index_f32u10__SVBool_tPKfu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_gather_s32index_f32(svbool_t pg, const float32_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_s32index_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4f32(<vscale x 4 x i1> [[PG]], float* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s32, index, _f32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_s64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_s64index_f64u10__SVBool_tPKdu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_gather_s64index_f64(svbool_t pg, const float64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_s64index_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.index.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, s64, index, _f64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_u32index_s32u10__SVBool_tPKiu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldff1_gather_u32index_s32(svbool_t pg, const int32_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_u32index_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u32, index, _s32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_u64index_s64u10__SVBool_tPKlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldff1_gather_u64index_s64(svbool_t pg, const int64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u64, index, _s64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_u32index_u32u10__SVBool_tPKju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldff1_gather_u32index_u32(svbool_t pg, const uint32_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_u32index_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u32, index, _u32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_u64index_u64u10__SVBool_tPKmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldff1_gather_u64index_u64(svbool_t pg, const uint64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u64, index, _u64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_u32index_f32u10__SVBool_tPKfu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldff1_gather_u32index_f32(svbool_t pg, const float32_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_u32index_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4f32(<vscale x 4 x i1> [[PG]], float* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u32, index, _f32)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldff1_gather_u64index_f64u10__SVBool_tPKdu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldff1_gather_u64index_f64(svbool_t pg, const float64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1_gather_u64index_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.index.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather_, u64, index, _f64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldff1_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldff1_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldff1_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldff1_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _index_u64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u32base_index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldff1_gather_u32base_index_f32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svldff1_gather_u32base_index_f32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1_gather_u32base_index_f32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u32base, _index_f32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1_gather_u64base_index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldff1_gather_u64base_index_f64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svldff1_gather_u64base_index_f64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1_gather_u64base_index_f64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   return SVE_ACLE_FUNC(svldff1_gather, _u64base, _index_f64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sb.c
index 745f97ca4adcb..7ebaf8455d011 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,278 +14,542 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1sb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sb_s16u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svldff1sb_s16(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1sb_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldff1sb_s16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sb_s32u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sb_s32(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1sb_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sb_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sb_s64u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sb_s64(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1sb_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sb_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sb_u16u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svldff1sb_u16(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1sb_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldff1sb_u16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sb_u32u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sb_u32(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1sb_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sb_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sb_u64u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sb_u64(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldff1sb_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sb_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sb_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sb_vnum_s16u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svint16_t test_svldff1sb_vnum_s16(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sb_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldff1sb_vnum_s16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sb_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sb_vnum_s32u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldff1sb_vnum_s32(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sb_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sb_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sb_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sb_vnum_s64u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldff1sb_vnum_s64(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sb_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sb_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sb_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sb_vnum_u16u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svuint16_t test_svldff1sb_vnum_u16(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sb_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldff1sb_vnum_u16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sb_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sb_vnum_u32u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldff1sb_vnum_u32(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sb_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sb_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sb_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sb_vnum_u64u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldff1sb_vnum_u64(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sb_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sb_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sb_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sb_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1sb_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sb_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sb_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1sb_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sb_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sb_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1sb_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sb_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sb_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1sb_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_s32offset_s32u10__SVBool_tPKau11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sb_gather_s32offset_s32(svbool_t pg, const int8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_s32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_s64offset_s64u10__SVBool_tPKau11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sb_gather_s64offset_s64(svbool_t pg, const int8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_s32offset_u32u10__SVBool_tPKau11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sb_gather_s32offset_u32(svbool_t pg, const int8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_s32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_s64offset_u64u10__SVBool_tPKau11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sb_gather_s64offset_u64(svbool_t pg, const int8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_u32offset_s32u10__SVBool_tPKau12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sb_gather_u32offset_s32(svbool_t pg, const int8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_u64offset_s64u10__SVBool_tPKau12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sb_gather_u64offset_s64(svbool_t pg, const int8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_u32offset_u32u10__SVBool_tPKau12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sb_gather_u32offset_u32(svbool_t pg, const int8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sb_gather_u64offset_u64u10__SVBool_tPKau12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sb_gather_u64offset_u64(svbool_t pg, const int8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sb_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sb_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sb_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sb_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sb_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sb_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sb_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sb_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sb_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sb_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sb_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sb_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sb_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sb_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sb_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sh.c
index 367ff5c7afb35..ecac83d5e2edb 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,346 +14,682 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1sh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sh_s32u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_s32(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldff1sh_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sh_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sh_s64u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_s64(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldff1sh_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sh_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sh_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sh_u32u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_u32(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldff1sh_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sh_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sh_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sh_u64u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_u64(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldff1sh_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sh_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sh_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sh_vnum_s32u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldff1sh_vnum_s32(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sh_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sh_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sh_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sh_vnum_s64u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldff1sh_vnum_s64(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sh_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sh_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sh_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sh_vnum_u32u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldff1sh_vnum_u32(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sh_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldff1sh_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sh_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sh_vnum_u64u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldff1sh_vnum_u64(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sh_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sh_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sh_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sh_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sh_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sh_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_s32offset_s32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_gather_s32offset_s32(svbool_t pg, const int16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_s32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_s64offset_s64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_gather_s64offset_s64(svbool_t pg, const int16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_s32offset_u32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_gather_s32offset_u32(svbool_t pg, const int16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_s32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_s64offset_u64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_gather_s64offset_u64(svbool_t pg, const int16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_u32offset_s32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_gather_u32offset_s32(svbool_t pg, const int16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_u64offset_s64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_gather_u64offset_s64(svbool_t pg, const int16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_u32offset_u32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_gather_u32offset_u32(svbool_t pg, const int16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sh_gather_u64offset_u64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_gather_u64offset_u64(svbool_t pg, const int16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sh_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sh_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sh_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sh_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_s32index_s32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_gather_s32index_s32(svbool_t pg, const int16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_s32index_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_s64index_s64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_gather_s64index_s64(svbool_t pg, const int16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_s32index_u32u10__SVBool_tPKsu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_gather_s32index_u32(svbool_t pg, const int16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_s32index_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_s64index_u64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_gather_s64index_u64(svbool_t pg, const int16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_u32index_s32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1sh_gather_u32index_s32(svbool_t pg, const int16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32index_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_u64index_s64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sh_gather_u64index_s64(svbool_t pg, const int16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_u32index_u32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1sh_gather_u32index_u32(svbool_t pg, const int16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32index_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sh_gather_u64index_u64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sh_gather_u64index_u64(svbool_t pg, const int16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1sh_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldff1sh_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1sh_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldff1sh_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1sh_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldff1sh_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1sh_gather_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1sh_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1sh_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldff1sh_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1sh_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sh_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sw.c
index 32d503344bc25..76b2b56513fef 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1sw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,174 +14,342 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1sw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sw_s64u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_s64(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldff1sw_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sw_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1sw_u64u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_u64(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldff1sw_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sw_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1sw_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sw_vnum_s64u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldff1sw_vnum_s64(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sw_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sw_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sw_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1sw_vnum_u64u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldff1sw_vnum_u64(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1sw_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldff1sw_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sw_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1sw_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sw_gather_s64offset_s64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_gather_s64offset_s64(svbool_t pg, const int32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sw_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sw_gather_s64offset_u64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_gather_s64offset_u64(svbool_t pg, const int32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sw_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sw_gather_u64offset_s64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_gather_u64offset_s64(svbool_t pg, const int32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1sw_gather_u64offset_u64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_gather_u64offset_u64(svbool_t pg, const int32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sw_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1sw_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sw_gather_s64index_s64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_gather_s64index_s64(svbool_t pg, const int32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1sw_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sw_gather_s64index_u64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_gather_s64index_u64(svbool_t pg, const int32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1sw_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sw_gather_u64index_s64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1sw_gather_u64index_s64(svbool_t pg, const int32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1sw_gather_u64index_u64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1sw_gather_u64index_u64(svbool_t pg, const int32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1sw_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldff1sw_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1sw_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1sw_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldff1sw_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1sw_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   return SVE_ACLE_FUNC(svldff1sw_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1ub.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1ub.c
index 3477177783c3c..2f1c56ac6c794 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1ub.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1ub.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,278 +14,542 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1ub_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1ub_s16u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svldff1ub_s16(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1ub_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldff1ub_s16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1ub_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1ub_s32u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1ub_s32(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1ub_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1ub_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1ub_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1ub_s64u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1ub_s64(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1ub_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1ub_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1ub_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1ub_u16u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svldff1ub_u16(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1ub_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldff1ub_u16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1ub_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1ub_u32u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1ub_u32(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1ub_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1ub_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1ub_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1ub_u64u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1ub_u64(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldff1ub_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1ub_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1ub_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1ub_vnum_s16u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svint16_t test_svldff1ub_vnum_s16(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1ub_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldff1ub_vnum_s16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1ub_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1ub_vnum_s32u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldff1ub_vnum_s32(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1ub_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1ub_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1ub_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1ub_vnum_s64u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldff1ub_vnum_s64(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1ub_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1ub_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1ub_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1ub_vnum_u16u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svuint16_t test_svldff1ub_vnum_u16(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1ub_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldff1ub_vnum_u16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1ub_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1ub_vnum_u32u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldff1ub_vnum_u32(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1ub_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1ub_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1ub_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1ub_vnum_u64u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldff1ub_vnum_u64(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1ub_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1ub_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1ub_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1ub_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1ub_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1ub_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1ub_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1ub_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1ub_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1ub_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1ub_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1ub_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1ub_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1ub_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_s32offset_s32u10__SVBool_tPKhu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1ub_gather_s32offset_s32(svbool_t pg, const uint8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_s32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_s64offset_s64u10__SVBool_tPKhu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1ub_gather_s64offset_s64(svbool_t pg, const uint8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_s32offset_u32u10__SVBool_tPKhu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1ub_gather_s32offset_u32(svbool_t pg, const uint8_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_s32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_s64offset_u64u10__SVBool_tPKhu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1ub_gather_s64offset_u64(svbool_t pg, const uint8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_u32offset_s32u10__SVBool_tPKhu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1ub_gather_u32offset_s32(svbool_t pg, const uint8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_u64offset_s64u10__SVBool_tPKhu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1ub_gather_u64offset_s64(svbool_t pg, const uint8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_u32offset_u32u10__SVBool_tPKhu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1ub_gather_u32offset_u32(svbool_t pg, const uint8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1ub_gather_u64offset_u64u10__SVBool_tPKhu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1ub_gather_u64offset_u64(svbool_t pg, const uint8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1ub_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1ub_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1ub_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1ub_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1ub_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1ub_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1ub_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1ub_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1ub_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1ub_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1ub_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1ub_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1ub_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1ub_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1ub_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uh.c
index 8c7eef1571928..5c7eb04af5fa7 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,346 +14,682 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1uh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1uh_s32u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_s32(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldff1uh_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1uh_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1uh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1uh_s64u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_s64(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldff1uh_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uh_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1uh_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1uh_u32u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_u32(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldff1uh_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1uh_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1uh_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1uh_u64u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_u64(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldff1uh_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uh_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1uh_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1uh_vnum_s32u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldff1uh_vnum_s32(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1uh_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1uh_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1uh_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1uh_vnum_s64u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldff1uh_vnum_s64(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1uh_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uh_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1uh_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1uh_vnum_u32u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldff1uh_vnum_u32(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1uh_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldff1uh_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1uh_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1uh_vnum_u64u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldff1uh_vnum_u64(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1uh_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uh_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1uh_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1uh_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1uh_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1uh_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_s32offset_s32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_gather_s32offset_s32(svbool_t pg, const uint16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_s32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_s64offset_s64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_gather_s64offset_s64(svbool_t pg, const uint16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_s32offset_u32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_gather_s32offset_u32(svbool_t pg, const uint16_t *base, svint32_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_s32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_s64offset_u64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_gather_s64offset_u64(svbool_t pg, const uint16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_u32offset_s32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_gather_u32offset_s32(svbool_t pg, const uint16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_u64offset_s64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_gather_u64offset_s64(svbool_t pg, const uint16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_u32offset_u32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_gather_u32offset_u32(svbool_t pg, const uint16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uh_gather_u64offset_u64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_gather_u64offset_u64(svbool_t pg, const uint16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1uh_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1uh_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1uh_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1uh_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_s32index_s32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_gather_s32index_s32(svbool_t pg, const uint16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_s32index_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_s64index_s64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_gather_s64index_s64(svbool_t pg, const uint16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_s32index_u32u10__SVBool_tPKtu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_gather_s32index_u32(svbool_t pg, const uint16_t *base, svint32_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_s32index_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.sxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_s64index_u64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_gather_s64index_u64(svbool_t pg, const uint16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_u32index_s32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldff1uh_gather_u32index_s32(svbool_t pg, const uint16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32index_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u32, index_s32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_u64index_s64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uh_gather_u64index_s64(svbool_t pg, const uint16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_u32index_u32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldff1uh_gather_u32index_u32(svbool_t pg, const uint16_t *base, svuint32_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32index_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.uxtw.index.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u32, index_u32, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uh_gather_u64index_u64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uh_gather_u64index_u64(svbool_t pg, const uint16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1uh_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldff1uh_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1uh_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldff1uh_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1uh_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldff1uh_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1uh_gather_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1uh_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1uh_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldff1uh_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1uh_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uh_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uw.c
index 8a38402f8af0c..7c9e08ed043aa 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1uw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,174 +14,342 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldff1uw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1uw_s64u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_s64(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldff1uw_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uw_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1uw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldff1uw_u64u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_u64(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldff1uw_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uw_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldff1uw_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1uw_vnum_s64u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldff1uw_vnum_s64(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1uw_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uw_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1uw_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldff1uw_vnum_u64u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldff1uw_vnum_u64(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldff1uw_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldff1uw_vnum_u64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1uw_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldff1uw_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uw_gather_s64offset_s64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_gather_s64offset_s64(svbool_t pg, const uint32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uw_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uw_gather_s64offset_u64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_gather_s64offset_u64(svbool_t pg, const uint32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uw_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uw_gather_u64offset_s64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_gather_u64offset_s64(svbool_t pg, const uint32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldff1uw_gather_u64offset_u64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_gather_u64offset_u64(svbool_t pg, const uint32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1uw_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldff1uw_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uw_gather_s64index_s64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_gather_s64index_s64(svbool_t pg, const uint32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1uw_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uw_gather_s64index_u64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_gather_s64index_u64(svbool_t pg, const uint32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldff1uw_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uw_gather_u64index_s64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldff1uw_gather_u64index_s64(svbool_t pg, const uint32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldff1uw_gather_u64index_u64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldff1uw_gather_u64index_u64(svbool_t pg, const uint32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1uw_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldff1uw_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldff1uw_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldff1uw_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldff1uw_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldff1uw_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldff1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   return SVE_ACLE_FUNC(svldff1uw_gather, _u64base, _index_u64, )(pg, bases, index);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c
index 005600a5de963..c0cc7a6a16c85 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 #include <arm_sve.h>
 
@@ -12,24 +13,42 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnf1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svldnf1_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svldnf1_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svldnf1_bf16'}}
   return SVE_ACLE_FUNC(svldnf1,_bf16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svldnf1_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
 svbfloat16_t test_svldnf1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_bf16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svldnf1_vnum_bf16'}}
   return SVE_ACLE_FUNC(svldnf1_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1.c
index b998b0b712c96..17628fa2f0a9a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,218 +14,412 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnf1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svldnf1_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svldnf1_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svldnf1_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldnf1_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnf1_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svldnf1_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svldnf1_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svldnf1_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldnf1_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnf1_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svldnf1_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnf1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnf1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldnf1_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnf1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnf1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnf1_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnf1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnf1_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svldnf1_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnf1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svldnf1_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svldnf1_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_s8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svldnf1_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_s16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldnf1_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_s32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldnf1_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_s64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svldnf1_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svldnf1_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_u8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnf1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svldnf1_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_u16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldnf1_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_u32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnf1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldnf1_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_u64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnf1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svldnf1_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_f16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnf1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnf1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svldnf1_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_f32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnf1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnf1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnf1_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnf1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svldnf1_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1_vnum_f64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnf1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnf1_vnum,_f64,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sb.c
index dd206157eff0e..242532999dcb0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sb.c
@@ -1,137 +1,258 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svldnf1sb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sb_s16u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svldnf1sb_s16(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sb_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldnf1sb_s16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sb_s32u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnf1sb_s32(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sb_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sb_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sb_s64u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnf1sb_s64(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sb_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sb_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sb_u16u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svldnf1sb_u16(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sb_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldnf1sb_u16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sb_u32u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnf1sb_u32(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sb_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sb_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sb_u64u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnf1sb_u64(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sb_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sb_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sb_vnum_s16u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svint16_t test_svldnf1sb_vnum_s16(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sb_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldnf1sb_vnum_s16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sb_vnum_s32u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldnf1sb_vnum_s32(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sb_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sb_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sb_vnum_s64u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldnf1sb_vnum_s64(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sb_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sb_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sb_vnum_u16u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svuint16_t test_svldnf1sb_vnum_u16(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sb_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[SEXT]]
   return svldnf1sb_vnum_u16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sb_vnum_u32u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldnf1sb_vnum_u32(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sb_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sb_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sb_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sb_vnum_u64u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldnf1sb_vnum_u64(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sb_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sb_vnum_u64(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sh.c
index 2d200c3117def..bfbdfd58475ce 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sh.c
@@ -1,93 +1,174 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svldnf1sh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sh_s32u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnf1sh_s32(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sh_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sh_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sh_s64u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnf1sh_s64(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sh_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sh_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sh_u32u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnf1sh_u32(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sh_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sh_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sh_u64u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnf1sh_u64(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sh_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sh_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sh_vnum_s32u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldnf1sh_vnum_s32(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sh_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sh_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sh_vnum_s64u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldnf1sh_vnum_s64(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sh_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sh_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sh_vnum_u32u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldnf1sh_vnum_u32(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sh_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[SEXT]]
   return svldnf1sh_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sh_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sh_vnum_u64u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldnf1sh_vnum_u64(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sh_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sh_vnum_u64(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sw.c
index 2c68efeb7a5f2..8eb3dcda9abfc 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1sw.c
@@ -1,49 +1,90 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svldnf1sw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sw_s64u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnf1sw_s64(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sw_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sw_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1sw_u64u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnf1sw_u64(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1sw_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sw_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1sw_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sw_vnum_s64u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldnf1sw_vnum_s64(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sw_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sw_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1sw_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1sw_vnum_u64u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = sext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldnf1sw_vnum_u64(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1sw_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[SEXT:.*]] = sext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[SEXT]]
   return svldnf1sw_vnum_u64(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1ub.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1ub.c
index 63407aa8cc345..005ba031b5fb4 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1ub.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1ub.c
@@ -1,137 +1,258 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svldnf1ub_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1ub_s16u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svldnf1ub_s16(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1ub_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldnf1ub_s16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1ub_s32u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnf1ub_s32(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1ub_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1ub_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1ub_s64u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnf1ub_s64(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1ub_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1ub_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1ub_u16u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 8 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svldnf1ub_u16(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1ub_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldnf1ub_u16(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1ub_u32u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnf1ub_u32(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1ub_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1ub_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1ub_u64u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnf1ub_u64(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnf1ub_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1ub_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1ub_vnum_s16u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svint16_t test_svldnf1ub_vnum_s16(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1ub_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldnf1ub_vnum_s16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1ub_vnum_s32u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldnf1ub_vnum_s32(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1ub_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1ub_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1ub_vnum_s64u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldnf1ub_vnum_s64(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1ub_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1ub_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1ub_vnum_u16u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 8 x i8> [[TMP3]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP4]]
+//
 svuint16_t test_svldnf1ub_vnum_u16(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1ub_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 8 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i8>, <vscale x 8 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 8 x i8> %[[LOAD]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[ZEXT]]
   return svldnf1ub_vnum_u16(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1ub_vnum_u32u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i8> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldnf1ub_vnum_u32(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1ub_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 4 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i8>, <vscale x 4 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i8> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1ub_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1ub_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1ub_vnum_u64u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i8>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i8> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldnf1ub_vnum_u64(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1ub_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 2 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i8>, <vscale x 2 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnf1.nxv2i8(<vscale x 2 x i1> %[[PG]], i8* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i8> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1ub_vnum_u64(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uh.c
index d60710759dd40..eaca5add89331 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uh.c
@@ -1,93 +1,174 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svldnf1uh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1uh_s32u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnf1uh_s32(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1uh_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1uh_s32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1uh_s64u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnf1uh_s64(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1uh_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uh_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1uh_u32u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnf1uh_u32(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1uh_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1uh_u32(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1uh_u64u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnf1uh_u64(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldnf1uh_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uh_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1uh_vnum_s32u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svint32_t test_svldnf1uh_vnum_s32(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1uh_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1uh_vnum_s32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1uh_vnum_s64u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldnf1uh_vnum_s64(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1uh_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uh_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1uh_vnum_u32u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 4 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 4 x i16> [[TMP3]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP4]]
+//
 svuint32_t test_svldnf1uh_vnum_u32(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1uh_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 4 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i16>, <vscale x 4 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 4 x i16> %[[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[ZEXT]]
   return svldnf1uh_vnum_u32(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1uh_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1uh_vnum_u64u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 2 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i16> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldnf1uh_vnum_u64(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1uh_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 2 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i16>, <vscale x 2 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnf1.nxv2i16(<vscale x 2 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i16> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uh_vnum_u64(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uw.c
index 9005268eb0b50..6f9d8b5080e22 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1uw.c
@@ -1,49 +1,90 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svldnf1uw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1uw_s64u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnf1uw_s64(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1uw_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uw_s64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1uw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svldnf1uw_u64u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnf1uw_u64(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldnf1uw_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %base)
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uw_u64(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnf1uw_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1uw_vnum_s64u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svint64_t test_svldnf1uw_vnum_s64(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1uw_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uw_vnum_s64(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnf1uw_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svldnf1uw_vnum_u64u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 2 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = zext <vscale x 2 x i32> [[TMP3]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP4]]
+//
 svuint64_t test_svldnf1uw_vnum_u64(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnf1uw_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 2 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i32>, <vscale x 2 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnf1.nxv2i32(<vscale x 2 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: %[[ZEXT:.*]] = zext <vscale x 2 x i32> %[[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[ZEXT]]
   return svldnf1uw_vnum_u64(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c
index 7d972a9dda84e..31755f85c56ab 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,24 +14,42 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svldnt1_bf16u10__SVBool_tPKu6__bf16(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svldnt1_bf16(svbool_t pg, const bfloat16_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_bf16'}}
   return SVE_ACLE_FUNC(svldnt1,_bf16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svldnt1_vnum_bf16u10__SVBool_tPKu6__bf16l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP3]]
+//
 svbfloat16_t test_svldnt1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_vnum_bf16'}}
   return SVE_ACLE_FUNC(svldnt1_vnum,_bf16,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1.c
index 35f9f6377b052..037bb7273138a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,218 +15,412 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svldnt1_s8u10__SVBool_tPKa(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svldnt1_s8(svbool_t pg, const int8_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_s8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_s8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_s16u10__SVBool_tPKs(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svldnt1_s16(svbool_t pg, const int16_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_s16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_s32u10__SVBool_tPKi(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldnt1_s32(svbool_t pg, const int32_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_s32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_s64u10__SVBool_tPKl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_s64(svbool_t pg, const int64_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_s64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svldnt1_u8u10__SVBool_tPKh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svldnt1_u8(svbool_t pg, const uint8_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_u8
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_u8,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_u16u10__SVBool_tPKt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svldnt1_u16(svbool_t pg, const uint16_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_u16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_u32u10__SVBool_tPKj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldnt1_u32(svbool_t pg, const uint32_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_u32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_u64u10__SVBool_tPKm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_u64(svbool_t pg, const uint64_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_u64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_f16u10__SVBool_tPKDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svldnt1_f16(svbool_t pg, const float16_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_f16,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_f32u10__SVBool_tPKf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldnt1_f32(svbool_t pg, const float32_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_f32,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svldnt1_f64u10__SVBool_tPKd(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_f64(svbool_t pg, const float64_t *base)
 {
-  // CHECK-LABEL: test_svldnt1_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1,_f64,,)(pg, base);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svldnt1_vnum_s8u10__SVBool_tPKal(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svldnt1_vnum_s8(svbool_t pg, const int8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_s8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_s8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_s16u10__SVBool_tPKsl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svldnt1_vnum_s16(svbool_t pg, const int16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_s16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_s32u10__SVBool_tPKil(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldnt1_vnum_s32(svbool_t pg, const int32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_s32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_s64u10__SVBool_tPKll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldnt1_vnum_s64(svbool_t pg, const int64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_s64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svldnt1_vnum_u8u10__SVBool_tPKhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svldnt1_vnum_u8(svbool_t pg, const uint8_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_u8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ldnt1.nxv16i8(<vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret <vscale x 16 x i8> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_u8,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_u16u10__SVBool_tPKtl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svldnt1_vnum_u16(svbool_t pg, const uint16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ldnt1.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret <vscale x 8 x i16> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_u16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_u32u10__SVBool_tPKjl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldnt1_vnum_u32(svbool_t pg, const uint32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret <vscale x 4 x i32> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_u32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_u64u10__SVBool_tPKml(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldnt1_vnum_u64(svbool_t pg, const uint64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret <vscale x 2 x i64> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_u64,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_f16u10__SVBool_tPKDhl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svldnt1_vnum_f16(svbool_t pg, const float16_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK: ret <vscale x 8 x half> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_f16,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_f32u10__SVBool_tPKfl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svldnt1_vnum_f32(svbool_t pg, const float32_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK: ret <vscale x 4 x float> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_f32,,)(pg, base, vnum);
 }
 
+// CHECK-LABEL: @test_svldnt1_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svldnt1_vnum_f64u10__SVBool_tPKdl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svldnt1_vnum_f64(svbool_t pg, const float64_t *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svldnt1_vnum_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: %[[LOAD:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK: ret <vscale x 2 x double> %[[LOAD]]
   return SVE_ACLE_FUNC(svldnt1_vnum,_f64,,)(pg, base, vnum);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len-bfloat.c
index ecba45608a7f1..c39ae28238355 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len-bfloat.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,12 +14,20 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlen_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlen_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_bf16(svbfloat16_t op)
 {
-  // CHECK-LABEL: test_svlen_bf16
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 3
-  // CHECK: ret i64 %[[SHL]]
   // expected-warning at +1 {{implicit declaration of function 'svlen_bf16'}}
   return SVE_ACLE_FUNC(svlen,_bf16,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len.c
index 115a176e18ae8..7b25b01c43900 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_len.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,101 +14,189 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlen_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svlen_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svlen_s8
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 4
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svlen_s16
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 3
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svlen_s32
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 2
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svlen_s64
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 1
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svlen_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 4
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svlen_u8
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 4
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svlen_u16
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 3
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svlen_u32
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 2
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svlen_u64
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 1
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 3
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svlen_f16
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 3
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 2
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svlen_f32
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 2
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svlen_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svlen_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[TMP0]], 1
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svlen_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svlen_f64
-  // CHECK: %[[VSCALE:.*]] = call i64 @llvm.vscale.i64()
-  // CHECK: %[[SHL:.*]] = shl i64 %[[VSCALE]], 1
-  // CHECK: ret i64 %[[SHL]]
   return SVE_ACLE_FUNC(svlen,_f64,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsl.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsl.c
index 9fd6c0d283038..cb212faa0c03a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsl.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsl.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,472 +14,891 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlsl_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsl_s8_zu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svlsl_s8_z(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s16_zu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svlsl_s16_z(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s32_zu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svlsl_s32_z(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s64_zu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svlsl_s64_z(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsl_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svlsl_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svlsl_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svlsl_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svlsl_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsl_s8_mu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svlsl_s8_m(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s16_mu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlsl_s16_m(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s32_mu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlsl_s32_m(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s64_mu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svlsl_s64_m(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsl_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsl_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsl_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsl_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svlsl_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsl_s8_xu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svlsl_s8_x(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s16_xu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlsl_s16_x(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s32_xu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlsl_s32_x(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_s64_xu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svlsl_s64_x(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsl_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsl_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsl_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsl_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsl_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svlsl_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsl_wide_s8_zu10__SVBool_tu10__SVInt8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svlsl_wide_s8_z(svbool_t pg, svint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_s16_zu10__SVBool_tu11__SVInt16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svlsl_wide_s16_z(svbool_t pg, svint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_s32_zu10__SVBool_tu11__SVInt32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svlsl_wide_s32_z(svbool_t pg, svint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsl_wide_u8_zu10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svlsl_wide_u8_z(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svlsl_wide_u16_z(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svlsl_wide_u32_z(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsl_wide_s8_mu10__SVBool_tu10__SVInt8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svlsl_wide_s8_m(svbool_t pg, svint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_s16_mu10__SVBool_tu11__SVInt16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlsl_wide_s16_m(svbool_t pg, svint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_s32_mu10__SVBool_tu11__SVInt32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlsl_wide_s32_m(svbool_t pg, svint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsl_wide_u8_mu10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsl_wide_u8_m(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsl_wide_u16_m(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsl_wide_u32_m(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsl_wide_s8_xu10__SVBool_tu10__SVInt8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svlsl_wide_s8_x(svbool_t pg, svint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_s16_xu10__SVBool_tu11__SVInt16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlsl_wide_s16_x(svbool_t pg, svint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_s32_xu10__SVBool_tu11__SVInt32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlsl_wide_s32_x(svbool_t pg, svint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsl_wide_u8_xu10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsl_wide_u8_x(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsl_wide_u16_x(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsl_wide_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsl_wide_u32_x(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svlsl_wide_n_s8_mu10__SVBool_tu10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svlsl_wide_n_s8_m(svbool_t pg, svint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsl_wide_n_s16_mu10__SVBool_tu11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svlsl_wide_n_s16_m(svbool_t pg, svint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsl_wide_n_s32_mu10__SVBool_tu11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svlsl_wide_n_s32_m(svbool_t pg, svint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svlsl_wide_n_s8_zu10__SVBool_tu10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svlsl_wide_n_s8_z(svbool_t pg, svint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[PG]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsl_wide_n_s16_zu10__SVBool_tu11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svlsl_wide_n_s16_z(svbool_t pg, svint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[OP]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsl_wide_n_s32_zu10__SVBool_tu11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svlsl_wide_n_s32_z(svbool_t pg, svint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[OP]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svlsl_wide_n_s8_xu10__SVBool_tu10__SVInt8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svlsl_wide_n_s8_x(svbool_t pg, svint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsl_wide_n_s16_xu10__SVBool_tu11__SVInt16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svlsl_wide_n_s16_x(svbool_t pg, svint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsl_wide_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsl_wide_n_s32_xu10__SVBool_tu11__SVInt32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svlsl_wide_n_s32_x(svbool_t pg, svint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsl_wide_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsl_wide,_n_s32,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsr.c
index 4357f1a87cd3f..3dd34eabf008d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_lsr.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,282 +14,532 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlsr_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsr_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svlsr_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svlsr_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svlsr_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svlsr_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsr_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsr_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsr_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsr_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svlsr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svlsr_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsr_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsr_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svlsr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svlsr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsr_wide_u8_zu10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svlsr_wide_u8_z(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsr_wide_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svlsr_wide_u16_z(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsr_wide_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svlsr_wide_u32_z(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsr_wide_u8_mu10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsr_wide_u8_m(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsr_wide_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsr_wide_u16_m(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsr_wide_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsr_wide_u32_m(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svlsr_wide_u8_xu10__SVBool_tu11__SVUint8_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svlsr_wide_u8_x(svbool_t pg, svuint8_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsr_wide_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svlsr_wide_u16_x(svbool_t pg, svuint16_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svlsr_wide_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svlsr_wide_u32_x(svbool_t pg, svuint32_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svlsr_wide_n_u8_mu10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svlsr_wide_n_u8_m(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsr_wide_n_u16_mu10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svlsr_wide_n_u16_m(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsr_wide_n_u32_mu10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svlsr_wide_n_u32_m(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svlsr_wide_n_u8_zu10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svlsr_wide_n_u8_z(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[PG]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsr_wide_n_u16_zu10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svlsr_wide_n_u16_z(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[OP]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsr_wide_n_u32_zu10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svlsr_wide_n_u32_z(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[OP]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svlsr_wide_n_u8_xu10__SVBool_tu11__SVUint8_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svlsr_wide_n_u8_x(svbool_t pg, svuint8_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsr_wide_n_u16_xu10__SVBool_tu12__SVUint16_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svlsr_wide_n_u16_x(svbool_t pg, svuint16_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.wide.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svlsr_wide_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svlsr_wide_n_u32_xu10__SVBool_tu12__SVUint32_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svlsr_wide_n_u32_x(svbool_t pg, svuint32_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svlsr_wide_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.wide.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svlsr_wide,_n_u32,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mad.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mad.c
index 49bd973c93d92..269d3918d0266 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mad.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mad.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,637 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmad_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmad_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmad_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmad_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmad_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmad_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmad_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmad_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmad_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmad_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmad_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmad_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmad_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmad_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmad_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmad_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmad_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmad_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmad_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmad_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmad_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmad_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmad_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmad_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmad_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmad_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmad_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmad_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmad_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmad_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmad_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmad_n_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmad_n_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmad_n_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmad_n_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmad_n_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmad_n_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmad_n_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmad_n_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmad_n_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmad_n_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmad_n_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmad_n_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmad_n_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmad_n_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmad_n_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmad_n_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmad_n_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmad_n_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmad_n_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmad_n_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmad_n_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmad_n_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmad_n_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmad_n_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmad_n_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmad_n_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmad_n_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mad.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmad_n_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mad.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmad_n_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mad.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmad_n_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mad.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmad_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmad_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmad_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmad_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmad_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmad_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmad_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmad_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmad_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmad_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmad_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmad_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmad_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmad_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmad_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmad_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmad_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmad_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmad_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmad_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmad_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmad_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmad_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmad_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmad_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmad_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmad_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmad_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmad_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmad_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmad_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmad,_n_f64,_x,)(pg, op1, op2, op3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp32.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp32.c
index 545c9f4d5d99d..4ced3aa2063a5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp32.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp32.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f32mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f32mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f32mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f32mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f32mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f32mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,9 +13,16 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmmla_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmmla.nxv4f32(<vscale x 4 x float> [[X:%.*]], <vscale x 4 x float> [[Y:%.*]], <vscale x 4 x float> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmmla_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmmla.nxv4f32(<vscale x 4 x float> [[X:%.*]], <vscale x 4 x float> [[Y:%.*]], <vscale x 4 x float> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmmla_f32(svfloat32_t x, svfloat32_t y, svfloat32_t z) {
-  // CHECK-LABEL: test_svmmla_f32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmmla.nxv4f32(<vscale x 4 x float> %x, <vscale x 4 x float> %y, <vscale x 4 x float> %z)
-  // CHECK: ret <vscale x 4 x float> %[[RET]]
   return SVE_ACLE_FUNC(svmmla,_f32,,)(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp64.c
index 1682a53ad4d8e..1a798c0cf5ba2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_matmul_fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,9 +13,16 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmmla_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmmla.nxv2f64(<vscale x 2 x double> [[X:%.*]], <vscale x 2 x double> [[Y:%.*]], <vscale x 2 x double> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmmla_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmmla.nxv2f64(<vscale x 2 x double> [[X:%.*]], <vscale x 2 x double> [[Y:%.*]], <vscale x 2 x double> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmmla_f64(svfloat64_t x, svfloat64_t y, svfloat64_t z) {
-  // CHECK-LABEL: test_svmmla_f64
-  // CHECK: %[[RET:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmmla.nxv2f64(<vscale x 2 x double> %x, <vscale x 2 x double> %y, <vscale x 2 x double> %z)
-  // CHECK: ret <vscale x 2 x double> %[[RET]]
   return SVE_ACLE_FUNC(svmmla,_f64,,)(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_max.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_max.c
index 246a025fa81ce..3ba13569c1789 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_max.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_max.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,638 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmax_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmax_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmax_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmax_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmax_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmax_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmax_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmax_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmax_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmax_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmax_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmax_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmax_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmax_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmax_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmax_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmax_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmax_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmax_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmax_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmax_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmax_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmax_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmax_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmax_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmax_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmax_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmax_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmax_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmax_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmax_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmax_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmax_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmax_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmax_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmax_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmax_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmax_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmax_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmax_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmax_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmax_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmax_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmax_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmax_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmax_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmax_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmax_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmax_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmax_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmax_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmax_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmax_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmax_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmax_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmax_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmax_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmax_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmax_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmax_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmax_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmax_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmax_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmax_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmax_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmax_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmax_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmax_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmax_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmax_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmax_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmax_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmax_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmax_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmax_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmax_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmax_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmax_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmax_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmax_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmax_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmax_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmax_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmax_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmax_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmax_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmax_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmax_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmax_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmax_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmax_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmax_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmax,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnm.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnm.c
index 881ebec77034d..3ff6c42b20e23 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnm.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnm.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,178 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmaxnm_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmaxnm_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmaxnm_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmaxnm_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmaxnm_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmaxnm_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmaxnm_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmaxnm_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmaxnm_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmaxnm_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmaxnm_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmaxnm_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmaxnm_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmaxnm_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmaxnm_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmaxnm_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmaxnm_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmaxnm_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmaxnm_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnm_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svmaxnm_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmaxnm_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnm_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnm,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnmv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnmv.c
index 933b46497745b..b9f1c9c04134f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnmv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxnmv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,29 +15,53 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmaxnmv_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fmaxnmv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxnmv_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fmaxnmv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svmaxnmv_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svmaxnmv_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.fmaxnmv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnmv,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxnmv_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fmaxnmv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxnmv_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fmaxnmv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svmaxnmv_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svmaxnmv_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.fmaxnmv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnmv,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxnmv_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fmaxnmv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxnmv_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fmaxnmv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svmaxnmv_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svmaxnmv_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.fmaxnmv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxnmv,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxv.c
index 85cd5329bf56a..e0dced690a597 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_maxv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,99 +15,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmaxv_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.smaxv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svmaxv_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.smaxv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svmaxv_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svmaxv_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.smaxv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.smaxv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.smaxv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svmaxv_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svmaxv_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.smaxv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.smaxv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.smaxv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svmaxv_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svmaxv_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.smaxv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.smaxv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.smaxv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svmaxv_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svmaxv_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.smaxv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.umaxv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svmaxv_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.umaxv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svmaxv_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svmaxv_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.umaxv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.umaxv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.umaxv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svmaxv_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svmaxv_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.umaxv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.umaxv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.umaxv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svmaxv_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svmaxv_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.umaxv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.umaxv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.umaxv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svmaxv_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svmaxv_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.umaxv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_u64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fmaxv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fmaxv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svmaxv_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svmaxv_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.fmaxv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fmaxv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fmaxv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svmaxv_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svmaxv_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.fmaxv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svmaxv_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fmaxv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmaxv_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fmaxv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svmaxv_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svmaxv_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.fmaxv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmaxv,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_min.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_min.c
index 0e124707c97a5..d4c15435ba1ba 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_min.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_min.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,638 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmin_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmin_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmin_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmin_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmin_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmin_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmin_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmin_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmin_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmin_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmin_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmin_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmin_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmin_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmin_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmin_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmin_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmin_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmin_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmin_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmin_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmin_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmin_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmin_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmin_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmin_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmin_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmin_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmin_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmin_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmin_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmin_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmin_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmin_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmin_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmin_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmin_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmin_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmin_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmin_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmin_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmin_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmin_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmin_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmin_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmin_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmin_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmin_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmin_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmin_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmin_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmin_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmin_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmin_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmin_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmin_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmin_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmin_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmin_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmin_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmin_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmin_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmin_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmin_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmin_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmin_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmin_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmin_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmin_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmin_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmin_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmin_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmin_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmin_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmin_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmin_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmin_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmin_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmin_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmin_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmin_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmin_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmin_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmin_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmin_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmin_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmin_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmin_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmin_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmin_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmin_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmin_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmin,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnm.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnm.c
index 6dbd5e89823dc..e1a353ea427e5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnm.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnm.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,178 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svminnm_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svminnm_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svminnm_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svminnm_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svminnm_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svminnm_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svminnm_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svminnm_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svminnm_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svminnm_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svminnm_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminnm_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svminnm_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svminnm_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svminnm_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svminnm_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svminnm_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svminnm_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svminnm_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svminnm_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnm_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svminnm_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svminnm_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svminnm_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnm,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnmv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnmv.c
index 43825d79cb302..7054ae9ed63a3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnmv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minnmv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,29 +15,53 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svminnmv_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fminnmv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminnmv_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fminnmv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svminnmv_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svminnmv_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.fminnmv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnmv,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminnmv_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fminnmv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminnmv_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fminnmv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svminnmv_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svminnmv_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.fminnmv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnmv,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminnmv_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fminnmv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminnmv_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fminnmv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svminnmv_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svminnmv_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.fminnmv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminnmv,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minv.c
index 248e3970806ea..2f3c04e3749dd 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_minv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,99 +15,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svminv_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.sminv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svminv_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.sminv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svminv_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svminv_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.sminv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.sminv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.sminv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svminv_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svminv_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.sminv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sminv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sminv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svminv_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svminv_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sminv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sminv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sminv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svminv_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svminv_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sminv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.uminv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svminv_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.uminv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svminv_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svminv_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.uminv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.uminv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.uminv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svminv_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svminv_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.uminv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uminv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uminv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svminv_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svminv_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uminv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uminv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uminv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svminv_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svminv_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uminv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_u64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fminv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret half [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_f16u10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call half @llvm.aarch64.sve.fminv.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret half [[TMP1]]
+//
 float16_t test_svminv_f16(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svminv_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call half @llvm.aarch64.sve.fminv.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret half %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_f16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fminv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret float [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_f32u10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call float @llvm.aarch64.sve.fminv.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret float [[TMP1]]
+//
 float32_t test_svminv_f32(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svminv_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call float @llvm.aarch64.sve.fminv.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret float %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_f32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svminv_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fminv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret double [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svminv_f64u10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call double @llvm.aarch64.sve.fminv.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret double [[TMP1]]
+//
 float64_t test_svminv_f64(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svminv_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call double @llvm.aarch64.sve.fminv.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret double %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svminv,_f64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mla.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mla.c
index cbf591b75ecc2..9d16676a7bb29 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mla.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mla.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,685 +14,1300 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmla_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmla_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmla_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmla_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmla_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmla_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmla_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmla_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmla_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmla_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmla_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmla_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmla_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmla_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmla_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmla_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmla_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmla_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmla_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmla_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmla_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmla_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmla_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmla_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmla_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmla_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmla_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmla_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmla_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmla_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmla_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmla_n_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmla_n_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmla_n_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmla_n_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmla_n_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmla_n_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmla_n_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmla_n_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmla_n_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmla_n_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmla_n_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmla_n_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmla_n_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmla_n_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmla_n_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmla_n_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmla_n_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmla_n_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmla_n_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmla_n_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmla_n_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmla_n_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmla_n_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmla_n_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmla_n_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmla_n_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmla_n_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmla_n_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmla_n_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmla_n_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmla_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmla_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmla_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmla_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmla_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmla_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmla_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmla_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmla_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmla_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmla_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmla_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmla_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmla_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmla_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmla_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmla_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmla_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmla_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmla_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmla_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmla_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmla_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmla_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmla_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmla_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmla_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmla_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmla_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmla_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla,_n_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmla_lane_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svmla_lane_f16(svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla_lane,_f16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_f16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_f16_1u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svmla_lane_f16_1(svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_f16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmla.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla_lane,_f16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmla_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmla_lane_f32(svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla_lane,_f32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_f32_1u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmla_lane_f32_1(svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmla.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla_lane,_f32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmla_lane_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmla_lane_f64(svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.lane.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla_lane,_f64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_f64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_f64_1u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmla_lane_f64_1(svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_f64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmla.lane.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmla_lane,_f64,,)(op1, op2, op3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mls.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mls.c
index 033b062dad712..badd6eca36600 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mls.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mls.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,685 +14,1300 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmls_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmls_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmls_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmls_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmls_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmls_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmls_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmls_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmls_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmls_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmls_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmls_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmls_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmls_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmls_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmls_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmls_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmls_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmls_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmls_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmls_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmls_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmls_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmls_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmls_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmls_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmls_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmls_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmls_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmls_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmls_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmls_n_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmls_n_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmls_n_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmls_n_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmls_n_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmls_n_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmls_n_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmls_n_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmls_n_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmls_n_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmls_n_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmls_n_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmls_n_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmls_n_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmls_n_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmls_n_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmls_n_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmls_n_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmls_n_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmls_n_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmls_n_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmls_n_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmls_n_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmls_n_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmls_n_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmls_n_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmls_n_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmls_n_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmls_n_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmls_n_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmls_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmls_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmls_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmls_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmls_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmls_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmls_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmls_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmls_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmls_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmls_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmls_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmls_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmls_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmls_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmls_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmls_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmls_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmls_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmls_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmls_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmls_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmls_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmls_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmls_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmls_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmls_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmls_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmls_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmls_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls,_n_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmls_lane_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_f16u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svmls_lane_f16(svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls_lane,_f16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_f16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_f16_1u13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svmls_lane_f16_1(svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_f16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmls.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls_lane,_f16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmls_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_f32u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmls_lane_f32(svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls_lane,_f32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_f32_1u13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmls_lane_f32_1(svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmls.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls_lane,_f32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmls_lane_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_f64u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmls_lane_f64(svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.lane.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls_lane,_f64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_f64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_f64_1u13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmls_lane_f64_1(svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_f64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmls.lane.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmls_lane,_f64,,)(op1, op2, op3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mmla.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mmla.c
index 6716fa3225aa5..273402e38e4c7 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mmla.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mmla.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,23 +13,44 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmmla_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smmla.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmmla_s32u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smmla.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmmla_s32(svint32_t x, svint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svmmla_s32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smmla.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svmmla,_s32,,)(x, y, z);
 }
 
+// CHECK-LABEL: @test_svmmla_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ummla.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmmla_u32u12__SVUint32_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ummla.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmmla_u32(svuint32_t x, svuint8_t y, svuint8_t z) {
-  // CHECK-LABEL: test_svmmla_u32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ummla.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svmmla,_u32,,)(x, y, z);
 }
 
+// CHECK-LABEL: @test_svusmmla_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usmmla.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svusmmla_s32u11__SVInt32_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usmmla.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svusmmla_s32(svint32_t x, svuint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svusmmla_s32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usmmla.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusmmla,_s32,,)(x, y, z);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mov.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mov.c
index b66525158052b..98cc79483be3c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mov.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mov.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmov_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.and.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]], <vscale x 16 x i1> [[OP]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svmov_b_zu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.and.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]], <vscale x 16 x i1> [[OP]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svmov_b_z(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svmov_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.and.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmov,_b,_z,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_msb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_msb.c
index 24c9deb86a507..3d90b3703bbfa 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_msb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_msb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,637 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmsb_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmsb_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmsb_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmsb_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmsb_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmsb_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmsb_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmsb_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmsb_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmsb_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmsb_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmsb_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmsb_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmsb_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmsb_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmsb_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmsb_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmsb_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmsb_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmsb_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmsb_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmsb_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmsb_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmsb_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmsb_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmsb_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmsb_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmsb_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmsb_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmsb_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmsb_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmsb_n_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmsb_n_s8_z(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmsb_n_s16_z(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmsb_n_s32_z(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmsb_n_s64_z(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmsb_n_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmsb_n_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u8,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmsb_n_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmsb_n_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmsb_n_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmsb_n_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmsb_n_s8_m(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmsb_n_s16_m(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmsb_n_s32_m(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmsb_n_s64_m(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmsb_n_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmsb_n_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u8,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmsb_n_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmsb_n_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmsb_n_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmsb_n_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmsb_n_s8_x(svbool_t pg, svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmsb_n_s16_x(svbool_t pg, svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmsb_n_s32_x(svbool_t pg, svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmsb_n_s64_x(svbool_t pg, svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_s64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmsb_n_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmsb_n_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.msb.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u8,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmsb_n_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-    // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.msb.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmsb_n_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.msb.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmsb_n_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.msb.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_u64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmsb_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmsb_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmsb_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmsb_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmsb_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmsb_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmsb_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmsb_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmsb_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmsb_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmsb_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmsb_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmsb_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmsb_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmsb_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmsb_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmsb_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmsb_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmsb_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmsb_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmsb_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svmsb_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmsb,_n_f64,_x,)(pg, op1, op2, op3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mul.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mul.c
index d2110cd1819b3..23d1a87891737 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mul.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mul.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,687 +14,1300 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmul_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmul_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmul_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmul_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmul_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmul_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmul_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmul_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmul_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmul_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmul_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmul_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmul_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmul_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmul_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmul_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmul_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmul_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmul_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmul_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmul_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmul_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmul_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmul_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmul_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmul_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmul_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmul_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmul_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmul_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmul_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmul_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmul_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmul_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmul_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmul_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmul_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmul_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmul_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmul_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmul_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmul_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmul_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmul_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmul_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmul_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmul_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmul_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmul_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmul_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmul_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmul_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmul_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmul_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmul_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmul_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmul_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmul_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmul_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmul_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmul_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmul_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmul_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmul_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmul_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmul_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmul_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmul_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmul_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmul_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmul_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmul_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmul_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmul_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmul_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmul_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmul_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmul_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmul_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmul_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmul_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmul_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmul_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmul_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmul_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmul_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmul_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmul_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmul_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmul_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmul_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul,_n_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmul_lane_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svmul_lane_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul_lane,_f16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_f16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_f16_1u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.lane.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svmul_lane_f16_1(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_f16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmul.lane.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 7)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul_lane,_f16,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmul_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmul_lane_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul_lane,_f32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_f32_1u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmul_lane_f32_1(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmul.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, i32 3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul_lane,_f32,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svmul_lane_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmul_lane_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.lane.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul_lane,_f64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_f64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_f64_1u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.lane.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svmul_lane_f64_1(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_f64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmul.lane.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmul_lane,_f64,,)(op1, op2, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulh.c
index 2614be69dec71..68827598b367f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,462 +14,874 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmulh_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmulh_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmulh_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmulh_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmulh_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmulh_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmulh_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmulh_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmulh_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmulh_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmulh_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmulh_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmulh_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmulh_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmulh_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmulh_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmulh_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmulh_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmulh_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmulh_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmulh_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmulh_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmulh_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmulh_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmulh_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmulh_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmulh_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmulh_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmulh_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmulh_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulh_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmulh_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmulh_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svmulh_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svmulh_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svmulh_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svmulh_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmulh_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svmulh_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svmulh_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svmulh_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svmulh_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmulh_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmulh_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmulh_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmulh_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmulh_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmulh_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmulh_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmulh_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmulh_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmulh_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmulh_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svmulh_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svmulh_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svmulh_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svmulh_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmulh_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svmulh_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svmulh_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svmulh_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulh_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulh_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svmulh_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svmulh_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulh,_n_u64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulx.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulx.c
index 1fb72675273cb..86a972b2505e8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulx.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_mulx.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,179 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmulx_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmulx_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmulx_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmulx_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmulx_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmulx_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmulx_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmulx_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmulx_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmulx_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmulx_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmulx_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svmulx_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svmulx_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svmulx_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmulx_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmulx_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmulx_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svmulx_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svmulx_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmulx_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmulx_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svmulx_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svmulx_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svmulx,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nand.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nand.c
index 95d3f906ea848..5c70b9daf7546 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nand.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nand.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnand_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nand.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnand_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nand.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svnand_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svnand_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nand.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnand,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_neg.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_neg.c
index 362fd4d9e30ba..16729b7dee9d5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_neg.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_neg.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,188 +14,353 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svneg_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svneg_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svneg_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svneg_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svneg_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svneg_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svneg_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svneg_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svneg_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svneg_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svneg_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svneg_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svneg_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svneg_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svneg_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svneg_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svneg_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svneg_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svneg_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svneg_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svneg_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svneg_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.neg.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svneg_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svneg_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.neg.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svneg_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svneg_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.neg.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svneg_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svneg_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.neg.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svneg_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svneg_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svneg_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svneg_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svneg_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svneg_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svneg_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svneg_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svneg_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svneg_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svneg_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svneg_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svneg_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svneg_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fneg.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svneg_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svneg_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fneg.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svneg_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svneg_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svneg_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svneg_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fneg.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svneg,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmad.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmad.c
index 80a8f19506dd3..5e708bdc2604d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmad.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmad.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,179 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnmad_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmad_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmad_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmad_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmad_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmad_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmad_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmad_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmad_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmad_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmad_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmad_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svnmad_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svnmad_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svnmad_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmad_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmad_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmad_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmad_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmad.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmad_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmad.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmad_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmad_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmad_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmad_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmad.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmad,_n_f64,_x,)(pg, op1, op2, op3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmla.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmla.c
index 8647fae29868e..a2ed49a4633b8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmla.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmla.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,179 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnmla_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmla_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmla_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmla_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmla_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmla_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmla_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmla_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmla_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmla_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmla_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmla_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svnmla_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svnmla_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svnmla_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmla_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmla_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmla_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmla_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmla_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmla_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmla_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmla_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmla_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmla,_n_f64,_x,)(pg, op1, op2, op3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmls.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmls.c
index 37547faba3d2f..c4698f529bdb2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmls.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmls.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,179 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnmls_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmls_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmls_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmls_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmls_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmls_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmls_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmls_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmls_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmls_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmls_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmls_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svnmls_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svnmls_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svnmls_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmls_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmls_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmls_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmls_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmls_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmls_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmls_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmls_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmls_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmls,_n_f64,_x,)(pg, op1, op2, op3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmsb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmsb.c
index e814a0925ccc9..c16c0cfe16b82 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmsb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nmsb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,179 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnmsb_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmsb_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmsb_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmsb_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmsb_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmsb_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmsb_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svnmsb_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svnmsb_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, svfloat32_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmsb_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svnmsb_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, svfloat64_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %op3)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_f64,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svnmsb_n_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f16,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svnmsb_n_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f32,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svnmsb_n_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f64,_z,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmsb_n_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f16,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmsb_n_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f32,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmsb_n_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f64,_m,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svnmsb_n_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fnmsb.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f16,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svnmsb_n_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2, float32_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fnmsb.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f32,_x,)(pg, op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnmsb_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svnmsb_n_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svnmsb_n_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2, float64_t op3)
 {
-  // CHECK-LABEL: test_svnmsb_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fnmsb.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnmsb,_n_f64,_x,)(pg, op1, op2, op3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nor.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nor.c
index bc274f3b20b78..f22cff4d36411 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nor.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_nor.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnor_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nor.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svnor_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nor.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svnor_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svnor_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nor.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnor,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_not.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_not.c
index 3e6c49f766a92..4b1fc06a33114 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_not.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_not.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,220 +14,413 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnot_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnot_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svnot_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svnot_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svnot_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svnot_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svnot_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svnot_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svnot_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svnot_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnot_u8_zu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svnot_u8_z(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svnot_u8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svnot_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svnot_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svnot_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svnot_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svnot_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svnot_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnot_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svnot_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svnot_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svnot_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svnot_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svnot_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svnot_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svnot_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svnot_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnot_u8_mu11__SVUint8_tu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svnot_u8_m(svuint8_t inactive, svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svnot_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svnot_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svnot_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svnot_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svnot_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svnot_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svnot_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnot_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svnot_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svnot_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svnot_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svnot_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svnot_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svnot_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svnot_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svnot_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnot_u8_xu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svnot_u8_x(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svnot_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.not.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svnot_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svnot_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.not.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svnot_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svnot_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.not.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnot_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svnot_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svnot_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.not.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_u64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svnot_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.eor.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]], <vscale x 16 x i1> [[PG]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svnot_b_zu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.eor.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]], <vscale x 16 x i1> [[PG]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svnot_b_z(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svnot_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.eor.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svnot,_b,_z,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orn.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orn.c
index 5970a3238e61e..ec71c9701096e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orn.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orn.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svorn_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.orn.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorn_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.orn.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svorn_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svorn_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.orn.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorn,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orr.c
index 3727afcd38080..2a8d656a78303 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orr.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,470 +14,889 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svorr_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svorr_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svorr_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svorr_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svorr_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svorr_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svorr_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svorr_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svorr_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svorr_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svorr_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svorr_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svorr_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svorr_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svorr_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svorr_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svorr_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svorr_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svorr_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svorr_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svorr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svorr_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svorr_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svorr_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svorr_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svorr_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svorr_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svorr_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svorr_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svorr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svorr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svorr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svorr_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svorr_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svorr_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svorr_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svorr_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svorr_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svorr_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svorr_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svorr_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svorr_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svorr_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svorr_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svorr_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svorr_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svorr_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svorr_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svorr_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svorr_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svorr_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svorr_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svorr_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svorr_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svorr_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svorr_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svorr_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svorr_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svorr_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svorr_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svorr_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svorr_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svorr_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svorr_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svorr_b_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.orr.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorr_b_zu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.orr.z.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svorr_b_z(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svorr_b_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.orr.z.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorr,_b,_z,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orv.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orv.c
index 6bb9cd0928ec2..6bbc6cc52e010 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orv.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_orv.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,72 +14,134 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svorv_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.orv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svorv_s8u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.orv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 int8_t test_svorv_s8(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svorv_s8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.orv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_s8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.orv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorv_s16u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.orv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 int16_t test_svorv_s16(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svorv_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.orv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_s16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.orv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorv_s32u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.orv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svorv_s32(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svorv_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.orv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_s32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.orv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorv_s64u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.orv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svorv_s64(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svorv_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.orv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_s64,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.orv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret i8 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svorv_u8u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i8 @llvm.aarch64.sve.orv.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i8 [[TMP0]]
+//
 uint8_t test_svorv_u8(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svorv_u8
-  // CHECK: %[[INTRINSIC:.*]] = call i8 @llvm.aarch64.sve.orv.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret i8 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_u8,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.orv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret i16 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorv_u16u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i16 @llvm.aarch64.sve.orv.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i16 [[TMP1]]
+//
 uint16_t test_svorv_u16(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svorv_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i16 @llvm.aarch64.sve.orv.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret i16 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_u16,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.orv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorv_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.orv.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svorv_u32(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svorv_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.orv.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_u32,,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svorv_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.orv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svorv_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.orv.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svorv_u64(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svorv_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.orv.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svorv,_u64,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfalse.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfalse.c
index 6d731d7176da7..cc871ad2120b5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfalse.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfalse.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,9 +14,15 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpfalse_b(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i1> zeroinitializer
+//
+// CPP-CHECK-LABEL: @_Z15test_svpfalse_bv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> zeroinitializer
+//
 svbool_t test_svpfalse_b()
 {
-  // CHECK-LABEL: test_svpfalse_b
-  // CHECK: ret <vscale x 16 x i1> zeroinitializer
   return SVE_ACLE_FUNC(svpfalse,_b,,)();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfirst.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfirst.c
index c69b406798f13..5376bf32ed93c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfirst.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pfirst.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpfirst_b(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.pfirst.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svpfirst_bu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.pfirst.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svpfirst_b(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svpfirst_b
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.pfirst.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svpfirst,_b,,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pnext.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pnext.c
index c9469845831a4..86996b20c551b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pnext.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_pnext.c
@@ -1,46 +1,84 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svpnext_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.pnext.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svpnext_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.pnext.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svpnext_b8(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svpnext_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.pnext.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svpnext_b8(pg, op);
 }
 
+// CHECK-LABEL: @test_svpnext_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.pnext.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svpnext_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.pnext.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svpnext_b16(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svpnext_b16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.pnext.nxv8i1(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i1> %[[OP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svpnext_b16(pg, op);
 }
 
+// CHECK-LABEL: @test_svpnext_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.pnext.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svpnext_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.pnext.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svpnext_b32(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svpnext_b32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.pnext.nxv4i1(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i1> %[[OP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svpnext_b32(pg, op);
 }
 
+// CHECK-LABEL: @test_svpnext_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.pnext.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svpnext_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.pnext.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svpnext_b64(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svpnext_b64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[OP:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.pnext.nxv2i1(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i1> %[[OP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svpnext_b64(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfb.c
index 3323eaf112ad4..cf133103df4e3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,168 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svprfb(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z11test_svprfbu10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 0)
   return svprfb(pg, base, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 1)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_1u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_1(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_1
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 1)
   return svprfb(pg, base, SV_PLDL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfb_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 2)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_2u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_2(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_2
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 2)
   return svprfb(pg, base, SV_PLDL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 3)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_3u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_3(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_3
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 3)
   return svprfb(pg, base, SV_PLDL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfb_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 4)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_4u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_4(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_4
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 4)
   return svprfb(pg, base, SV_PLDL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 5)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_5u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 5)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_5(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_5
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 5)
   return svprfb(pg, base, SV_PLDL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfb_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 8)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_6u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_6(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_6
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 8)
   return svprfb(pg, base, SV_PSTL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 9)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_7u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 9)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_7(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_7
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 9)
   return svprfb(pg, base, SV_PSTL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfb_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 10)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_8u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 10)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_8(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_8
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 10)
   return svprfb(pg, base, SV_PSTL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 11)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfb_9u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 11)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_9(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_9
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 11)
   return svprfb(pg, base, SV_PSTL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfb_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 12)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfb_10u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 12)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_10(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_10
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 12)
   return svprfb(pg, base, SV_PSTL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 13)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfb_11u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]], i32 13)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_11(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfb_11
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %base, i32 13)
   return svprfb(pg, base, SV_PSTL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfb_vnum(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svprfb_vnumu10__SVBool_tPKvl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_vnum(svbool_t pg, const void *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svprfb_vnum
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: @llvm.aarch64.sve.prf.nxv16i1(<vscale x 16 x i1> %pg, i8* %[[GEP]], i32 0)
   return svprfb_vnum(pg, base, vnum, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_u32base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfb_gather_u32baseu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_u32base(svbool_t pg, svuint32_t bases)
 {
-  // CHECK-LABEL: test_svprfb_gather_u32base
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather,_u32base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_u64base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfb_gather_u64baseu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_u64base(svbool_t pg, svuint64_t bases)
 {
-  // CHECK-LABEL: test_svprfb_gather_u64base
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather,_u64base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_s32offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z28test_svprfb_gather_s32offsetu10__SVBool_tPKvu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_s32offset(svbool_t pg, const void *base, svint32_t offsets)
 {
-  // CHECK-LABEL: test_svprfb_gather_s32offset
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather_,s32,offset,)(pg, base, offsets, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_s64offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z28test_svprfb_gather_s64offsetu10__SVBool_tPKvu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_s64offset(svbool_t pg, const void *base, svint64_t offsets)
 {
-  // CHECK-LABEL: test_svprfb_gather_s64offset
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather_,s64,offset,)(pg, base, offsets, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_u32offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z28test_svprfb_gather_u32offsetu10__SVBool_tPKvu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_u32offset(svbool_t pg, const void *base, svuint32_t offsets)
 {
-  // CHECK-LABEL: test_svprfb_gather_u32offset
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %offsets, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather_,u32,offset,)(pg, base, offsets, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_u64offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z28test_svprfb_gather_u64offsetu10__SVBool_tPKvu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_u64offset(svbool_t pg, const void *base, svuint64_t offsets)
 {
-  // CHECK-LABEL: test_svprfb_gather_u64offset
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %offsets, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather_,u64,offset,)(pg, base, offsets, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_u32base_offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svprfb_gather_u32base_offsetu10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_u32base_offset(svbool_t pg, svuint32_t bases, int64_t offset)
 {
-  // CHECK-LABEL: test_svprfb_gather_u32base_offset
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset, i32 0)
-  // CHECK: ret void
   return svprfb_gather_u32base_offset(pg, bases, offset, SV_PLDL1KEEP);
   return SVE_ACLE_FUNC(svprfb_gather,_u32base,_offset,)(pg, bases, offset, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfb_gather_u64base_offset(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svprfb_gather_u64base_offsetu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfb_gather_u64base_offset(svbool_t pg, svuint64_t bases, int64_t offset)
 {
-  // CHECK-LABEL: test_svprfb_gather_u64base_offset
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfb.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfb_gather,_u64base,_offset,)(pg, bases, offset, SV_PLDL1KEEP);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfd.c
index 9c331696998c9..5a71299f1b707 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,183 +14,369 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svprfd(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z11test_svprfdu10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 0)
   return svprfd(pg, base, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 1)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_1u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_1(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 1)
   return svprfd(pg, base, SV_PLDL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfd_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 2)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_2u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_2(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_2
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 2)
   return svprfd(pg, base, SV_PLDL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 3)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_3u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_3(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_3
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 3)
   return svprfd(pg, base, SV_PLDL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfd_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 4)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_4u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_4(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_4
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 4)
   return svprfd(pg, base, SV_PLDL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 5)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_5u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 5)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_5(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_5
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 5)
   return svprfd(pg, base, SV_PLDL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfd_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 8)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_6u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_6(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_6
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 8)
   return svprfd(pg, base, SV_PSTL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 9)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_7u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 9)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_7(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_7
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 9)
   return svprfd(pg, base, SV_PSTL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfd_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 10)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_8u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 10)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_8(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_8
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 10)
   return svprfd(pg, base, SV_PSTL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 11)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfd_9u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 11)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_9(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_9
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 11)
   return svprfd(pg, base, SV_PSTL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfd_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 12)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfd_10u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 12)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_10(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_10
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 12)
   return svprfd(pg, base, SV_PSTL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 13)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfd_11u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 13)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_11(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfd_11
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %base, i32 13)
   return svprfd(pg, base, SV_PSTL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfd_vnum(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = bitcast <vscale x 2 x i64>* [[TMP2]] to i8*
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[TMP3]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svprfd_vnumu10__SVBool_tPKvl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]]
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = bitcast <vscale x 2 x i64>* [[TMP2]] to i8*
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> [[TMP0]], i8* [[TMP3]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_vnum(svbool_t pg, const void *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svprfd_vnum
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum
-  // CHECK-DAG: %[[I8_BASE:.*]] = bitcast <vscale x 2 x i64>* %[[GEP]] to i8*
-  // CHECK: @llvm.aarch64.sve.prf.nxv2i1(<vscale x 2 x i1> %[[PG]], i8* %[[I8_BASE]], i32 0)
   return svprfd_vnum(pg, base, vnum, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_u32base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfd_gather_u32baseu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_u32base(svbool_t pg, svuint32_t bases)
 {
-  // CHECK-LABEL: test_svprfd_gather_u32base
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather,_u32base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_u64base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfd_gather_u64baseu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_u64base(svbool_t pg, svuint64_t bases)
 {
-  // CHECK-LABEL: test_svprfd_gather_u64base
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather,_u64base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_s32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfd_gather_s32indexu10__SVBool_tPKvu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_s32index(svbool_t pg, const void *base, svint32_t indices)
 {
-  // CHECK-LABEL: test_svprfd_gather_s32index
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather_,s32,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_s64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfd_gather_s64indexu10__SVBool_tPKvu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_s64index(svbool_t pg, const void *base, svint64_t indices)
 {
-  // CHECK-LABEL: test_svprfd_gather_s64index
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather_,s64,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_u32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfd_gather_u32indexu10__SVBool_tPKvu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_u32index(svbool_t pg, const void *base, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svprfd_gather_u32index
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather_,u32,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_u64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfd_gather_u64indexu10__SVBool_tPKvu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_u64index(svbool_t pg, const void *base, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svprfd_gather_u64index
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather_,u64,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_u32base_index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svprfd_gather_u32base_indexu10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_u32base_index(svbool_t pg, svuint32_t bases, int64_t index)
 {
-  // CHECK-LABEL: test_svprfd_gather_u32base_index
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]], i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather,_u32base,_index,)(pg, bases, index, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfd_gather_u64base_index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svprfd_gather_u64base_indexu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfd_gather_u64base_index(svbool_t pg, svuint64_t bases, int64_t index)
 {
-  // CHECK-LABEL: test_svprfd_gather_u64base_index
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.prfd.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]], i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfd_gather,_u64base,_index,)(pg, bases, index, SV_PLDL1KEEP);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfh.c
index ac4c8952e93d4..64739a942b3b0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,181 +14,369 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svprfh(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z11test_svprfhu10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 0)
   return svprfh(pg, base, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 1)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_1u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_1(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 1)
   return svprfh(pg, base, SV_PLDL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfh_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 2)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_2u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_2(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_2
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 2)
   return svprfh(pg, base, SV_PLDL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 3)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_3u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_3(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_3
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 3)
   return svprfh(pg, base, SV_PLDL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfh_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 4)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_4u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_4(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_4
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 4)
   return svprfh(pg, base, SV_PLDL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 5)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_5u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 5)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_5(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_5
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 5)
   return svprfh(pg, base, SV_PLDL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfh_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 8)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_6u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_6(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_6
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 8)
   return svprfh(pg, base, SV_PSTL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 9)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_7u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 9)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_7(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_7
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 9)
   return svprfh(pg, base, SV_PSTL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfh_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 10)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_8u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 10)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_8(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_8
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 10)
   return svprfh(pg, base, SV_PSTL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 11)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfh_9u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 11)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_9(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_9
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 11)
   return svprfh(pg, base, SV_PSTL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfh_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 12)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfh_10u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 12)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_10(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_10
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 12)
   return svprfh(pg, base, SV_PSTL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 13)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfh_11u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 13)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_11(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfh_11
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %base, i32 13)
   return svprfh(pg, base, SV_PSTL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfh_vnum(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = bitcast <vscale x 8 x i16>* [[TMP2]] to i8*
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[TMP3]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svprfh_vnumu10__SVBool_tPKvl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]]
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = bitcast <vscale x 8 x i16>* [[TMP2]] to i8*
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> [[TMP0]], i8* [[TMP3]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_vnum(svbool_t pg, const void *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svprfh_vnum
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum
-  // CHECK-DAG: %[[I8_BASE:.*]] = bitcast <vscale x 8 x i16>* %[[GEP]] to i8*
-  // CHECK: @llvm.aarch64.sve.prf.nxv8i1(<vscale x 8 x i1> %[[PG]], i8* %[[I8_BASE]], i32 0)
   return svprfh_vnum(pg, base, vnum, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_u32base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfh_gather_u32baseu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_u32base(svbool_t pg, svuint32_t bases)
 {
-  // CHECK-LABEL: test_svprfh_gather_u32base
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfh_gather,_u32base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_u64base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfh_gather_u64baseu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_u64base(svbool_t pg, svuint64_t bases)
 {
-  // CHECK-LABEL: test_svprfh_gather_u64base
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0, i32 0)
   return SVE_ACLE_FUNC(svprfh_gather,_u64base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_s32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfh_gather_s32indexu10__SVBool_tPKvu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_s32index(svbool_t pg, const void *base, svint32_t indices)
 {
-  // CHECK-LABEL: test_svprfh_gather_s32index
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfh_gather_,s32,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_s64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfh_gather_s64indexu10__SVBool_tPKvu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_s64index(svbool_t pg, const void *base, svint64_t indices)
 {
-  // CHECK-LABEL: test_svprfh_gather_s64index
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %indices, i32 0)
   return SVE_ACLE_FUNC(svprfh_gather_,s64,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_u32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfh_gather_u32indexu10__SVBool_tPKvu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_u32index(svbool_t pg, const void *base, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svprfh_gather_u32index
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfh_gather_,u32,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_u64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfh_gather_u64indexu10__SVBool_tPKvu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_u64index(svbool_t pg, const void *base, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svprfh_gather_u64index
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfh_gather_,u64,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_u32base_index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svprfh_gather_u32base_indexu10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_u32base_index(svbool_t pg, svuint32_t bases, int64_t index)
 {
-  // CHECK-LABEL: test_svprfh_gather_u32base_index
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]], i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfh_gather,_u32base,_index,)(pg, bases, index, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfh_gather_u64base_index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svprfh_gather_u64base_indexu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfh_gather_u64base_index(svbool_t pg, svuint64_t bases, int64_t index)
 {
-  // CHECK-LABEL: test_svprfh_gather_u64base_index
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 1
-  // CHECK: call void @llvm.aarch64.sve.prfh.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]], i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfh_gather,_u64base,_index,)(pg, bases, index, SV_PLDL1KEEP);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfw.c
index 223233aeb6cf6..8850f85a4772e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_prfw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,183 +14,369 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svprfw(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z11test_svprfwu10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 0)
   return svprfw(pg, base, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 1)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_1u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_1(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 1)
   return svprfw(pg, base, SV_PLDL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfw_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 2)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_2u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_2(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_2
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 2)
   return svprfw(pg, base, SV_PLDL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 3)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_3u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_3(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_3
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 3)
   return svprfw(pg, base, SV_PLDL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfw_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 4)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_4u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_4(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_4
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 4)
   return svprfw(pg, base, SV_PLDL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 5)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_5u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 5)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_5(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_5
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 5)
   return svprfw(pg, base, SV_PLDL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfw_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 8)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_6u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_6(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_6
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 8)
   return svprfw(pg, base, SV_PSTL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 9)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_7u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 9)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_7(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_7
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 9)
   return svprfw(pg, base, SV_PSTL1STRM);
 }
 
+// CHECK-LABEL: @test_svprfw_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 10)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_8u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 10)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_8(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_8
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 10)
   return svprfw(pg, base, SV_PSTL2KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 11)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svprfw_9u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 11)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_9(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_9
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 11)
   return svprfw(pg, base, SV_PSTL2STRM);
 }
 
+// CHECK-LABEL: @test_svprfw_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 12)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfw_10u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 12)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_10(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_10
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 12)
   return svprfw(pg, base, SV_PSTL3KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 13)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svprfw_11u10__SVBool_tPKv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], i32 13)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_11(svbool_t pg, const void *base)
 {
-  // CHECK-LABEL: test_svprfw_11
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %base, i32 13)
   return svprfw(pg, base, SV_PSTL3STRM);
 }
 
+// CHECK-LABEL: @test_svprfw_vnum(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]]
+// CHECK-NEXT:    [[TMP3:%.*]] = bitcast <vscale x 4 x i32>* [[TMP2]] to i8*
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[TMP3]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svprfw_vnumu10__SVBool_tPKvl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]]
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = bitcast <vscale x 4 x i32>* [[TMP2]] to i8*
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> [[TMP0]], i8* [[TMP3]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_vnum(svbool_t pg, const void *base, int64_t vnum)
 {
-  // CHECK-LABEL: test_svprfw_vnum
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i8* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum
-  // CHECK-DAG: %[[I8_BASE:.*]] = bitcast <vscale x 4 x i32>* %[[GEP]] to i8*
-  // CHECK: @llvm.aarch64.sve.prf.nxv4i1(<vscale x 4 x i1> %[[PG]], i8* %[[I8_BASE]], i32 0)
   return svprfw_vnum(pg, base, vnum, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_u32base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfw_gather_u32baseu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_u32base(svbool_t pg, svuint32_t bases)
 {
-  // CHECK-LABEL: test_svprfw_gather_u32base
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather,_u32base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_u64base(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z26test_svprfw_gather_u64baseu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0, i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_u64base(svbool_t pg, svuint64_t bases)
 {
-  // CHECK-LABEL: test_svprfw_gather_u64base
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather,_u64base,,)(pg, bases, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_s32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfw_gather_s32indexu10__SVBool_tPKvu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_s32index(svbool_t pg, const void *base, svint32_t indices)
 {
-  // CHECK-LABEL: test_svprfw_gather_s32index
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.sxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather_,s32,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_s64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfw_gather_s64indexu10__SVBool_tPKvu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_s64index(svbool_t pg, const void *base, svint64_t indices)
 {
-  // CHECK-LABEL: test_svprfw_gather_s64index
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather_,s64,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_u32index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfw_gather_u32indexu10__SVBool_tPKvu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_u32index(svbool_t pg, const void *base, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svprfw_gather_u32index
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.uxtw.index.nxv4i32(<vscale x 4 x i1> %[[PG]], i8* %base, <vscale x 4 x i32> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather_,u32,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_u64index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z27test_svprfw_gather_u64indexu10__SVBool_tPKvu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_u64index(svbool_t pg, const void *base, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svprfw_gather_u64index
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.index.nxv2i64(<vscale x 2 x i1> %[[PG]], i8* %base, <vscale x 2 x i64> %indices, i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather_,u64,index,)(pg, base, indices, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_u32base_index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svprfw_gather_u32base_indexu10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_u32base_index(svbool_t pg, svuint32_t bases, int64_t index)
 {
-  // CHECK-LABEL: test_svprfw_gather_u32base_index
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]], i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather,_u32base,_index,)(pg, bases, index, SV_PLDL1KEEP);
 }
 
+// CHECK-LABEL: @test_svprfw_gather_u64base_index(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svprfw_gather_u64base_indexu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svprfw_gather_u64base_index(svbool_t pg, svuint64_t bases, int64_t index)
 {
-  // CHECK-LABEL: test_svprfw_gather_u64base_index
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.prfw.gather.scalar.offset.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]], i32 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svprfw_gather,_u64base,_index,)(pg, bases, index, SV_PLDL1KEEP);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptest.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptest.c
index faa37b93fbfa3..8d6ec123a3053 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptest.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptest.c
@@ -1,29 +1,51 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svptest_any(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i1 @llvm.aarch64.sve.ptest.any.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret i1 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svptest_anyu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i1 @llvm.aarch64.sve.ptest.any.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i1 [[TMP0]]
+//
 bool test_svptest_any(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svptest_any
-  // CHECK: %[[INTRINSIC:.*]] = call i1 @llvm.aarch64.sve.ptest.any{{(.nxv16i1)?}}(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret i1 %[[INTRINSIC]]
   return svptest_any(pg, op);
 }
 
+// CHECK-LABEL: @test_svptest_first(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i1 @llvm.aarch64.sve.ptest.first.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret i1 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svptest_firstu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i1 @llvm.aarch64.sve.ptest.first.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i1 [[TMP0]]
+//
 bool test_svptest_first(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svptest_first
-  // CHECK: %[[INTRINSIC:.*]] = call i1 @llvm.aarch64.sve.ptest.first{{(.nxv16i1)?}}(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret i1 %[[INTRINSIC]]
   return svptest_first(pg, op);
 }
 
+// CHECK-LABEL: @test_svptest_last(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i1 @llvm.aarch64.sve.ptest.last.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret i1 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svptest_lastu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i1 @llvm.aarch64.sve.ptest.last.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret i1 [[TMP0]]
+//
 bool test_svptest_last(svbool_t pg, svbool_t op)
 {
-  // CHECK-LABEL: test_svptest_last
-  // CHECK: %[[INTRINSIC:.*]] = call i1 @llvm.aarch64.sve.ptest.last{{(.nxv16i1)?}}(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op)
-  // CHECK: ret i1 %[[INTRINSIC]]
   return svptest_last(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptrue.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptrue.c
index 70f9dbf3bb2f0..2870614843587 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptrue.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ptrue.c
@@ -1,203 +1,378 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svptrue_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svptrue_b8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_b8()
 {
-  // CHECK-LABEL: test_svptrue_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_b8();
 }
 
+// CHECK-LABEL: @test_svptrue_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svptrue_b16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svptrue_b16()
 {
-  // CHECK-LABEL: test_svptrue_b16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 31)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svptrue_b16();
 }
 
+// CHECK-LABEL: @test_svptrue_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svptrue_b32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svptrue_b32()
 {
-  // CHECK-LABEL: test_svptrue_b32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 31)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svptrue_b32();
 }
 
+// CHECK-LABEL: @test_svptrue_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svptrue_b64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svptrue_b64()
 {
-  // CHECK-LABEL: test_svptrue_b64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 31)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svptrue_b64();
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svptrue_pat_b8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 0)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_POW2);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_1v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_1()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 1)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL1);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 2)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_2v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_2()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL2);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 3)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_3v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_3()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 3)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL3);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 4)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_4v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 4)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_4()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_4
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 4)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL4);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_5(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 5)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_5v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 5)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_5()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_5
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 5)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL5);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_6(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 6)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_6v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 6)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_6()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_6
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 6)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL6);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_7(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 7)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_7v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_7()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_7
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 7)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL7);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_8()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 8)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL8);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_9(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 9)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svptrue_pat_b8_9v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 9)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_9()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_9
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 9)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL16);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_10(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 10)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_10v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 10)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_10()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_10
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 10)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL32);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_11(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 11)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_11v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 11)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_11()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_11
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 11)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL64);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_12(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 12)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_12v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 12)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_12()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_12
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 12)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL128);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_13(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 13)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_13v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 13)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_13()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_13
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 13)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_VL256);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_14(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 29)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_14v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 29)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_14()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_14
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 29)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_MUL4);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_15(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 30)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_15v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 30)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_15()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_15
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 30)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_MUL3);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b8_16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svptrue_pat_b8_16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svptrue_pat_b8_16()
 {
-  // CHECK-LABEL: test_svptrue_pat_b8_16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svptrue_pat_b8(SV_ALL);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svptrue_pat_b16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svptrue_pat_b16()
 {
-  // CHECK-LABEL: test_svptrue_pat_b16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.ptrue.nxv8i1(i32 0)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svptrue_pat_b16(SV_POW2);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 1)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svptrue_pat_b32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 1)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svptrue_pat_b32()
 {
-  // CHECK-LABEL: test_svptrue_pat_b32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.ptrue.nxv4i1(i32 1)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svptrue_pat_b32(SV_VL1);
 }
 
+// CHECK-LABEL: @test_svptrue_pat_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 2)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svptrue_pat_b64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 2)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svptrue_pat_b64()
 {
-  // CHECK-LABEL: test_svptrue_pat_b64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.ptrue.nxv2i1(i32 2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svptrue_pat_b64(SV_VL2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qadd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qadd.c
index 3cb66cc0b85a8..08e0c071952a6 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qadd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qadd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,138 +14,258 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqadd_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svqadd_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqadd_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqadd_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqadd_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqadd_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqadd_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqadd_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqadd_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svqadd_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqadd_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqadd_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqadd_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqadd_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqadd_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqadd_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svqadd_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_n_s8u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqadd_n_s8(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqadd_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqadd_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqadd_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqadd_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqadd_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqadd_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqadd_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqadd,_n_u64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecb.c
index f6c3dbb201ecf..b39e51d016e69 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,74 +14,137 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdecb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecb_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecb_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecb_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecb_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecb_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecb_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdecb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecb.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecb_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecb_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdecb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecb.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecb_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecb_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdecb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecb.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecb_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 [[OP:%.*]], i32 0, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecb_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 [[OP:%.*]], i32 0, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecb_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecb_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecb.n32(i32 %op, i32 0, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb_pat,_n_s32,,)(op, SV_POW2, 16);
 }
 
+// CHECK-LABEL: @test_svqdecb_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecb.n64(i64 [[OP:%.*]], i32 1, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecb_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecb.n64(i64 [[OP:%.*]], i32 1, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecb_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdecb_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecb.n64(i64 %op, i32 1, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb_pat,_n_s64,,)(op, SV_VL1, 1);
 }
 
+// CHECK-LABEL: @test_svqdecb_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecb.n32(i32 [[OP:%.*]], i32 2, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecb_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecb.n32(i32 [[OP:%.*]], i32 2, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecb_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdecb_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecb.n32(i32 %op, i32 2, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb_pat,_n_u32,,)(op, SV_VL2, 16);
 }
 
+// CHECK-LABEL: @test_svqdecb_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecb.n64(i64 [[OP:%.*]], i32 3, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecb_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecb.n64(i64 [[OP:%.*]], i32 3, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecb_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdecb_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecb.n64(i64 %op, i32 3, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecb_pat,_n_u64,,)(op, SV_VL3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecd.c
index fa9def9eb39ea..c948a17595dd0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,106 +14,197 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdecd_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecd_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecd_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecd_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecd_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecd_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecd_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecd_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecd_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecd_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecd_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecd.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecd_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecd_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecd_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdecd_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecd.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecd_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecd_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecd_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecd.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecd_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 [[OP:%.*]], i32 4, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecd_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 [[OP:%.*]], i32 4, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecd_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecd_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecd.n32(i32 %op, i32 4, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd_pat,_n_s32,,)(op, SV_VL4, 16);
 }
 
+// CHECK-LABEL: @test_svqdecd_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecd.n64(i64 [[OP:%.*]], i32 5, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecd_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecd.n64(i64 [[OP:%.*]], i32 5, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecd_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecd.n64(i64 %op, i32 5, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd_pat,_n_s64,,)(op, SV_VL5, 1);
 }
 
+// CHECK-LABEL: @test_svqdecd_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecd.n32(i32 [[OP:%.*]], i32 6, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecd_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecd.n32(i32 [[OP:%.*]], i32 6, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecd_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdecd_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecd.n32(i32 %op, i32 6, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd_pat,_n_u32,,)(op, SV_VL6, 16);
 }
 
+// CHECK-LABEL: @test_svqdecd_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecd.n64(i64 [[OP:%.*]], i32 7, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecd_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecd.n64(i64 [[OP:%.*]], i32 7, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecd_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecd.n64(i64 %op, i32 7, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd_pat,_n_u64,,)(op, SV_VL7, 1);
 }
 
+// CHECK-LABEL: @test_svqdecd_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecd_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdecd_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecd.nxv2i64(<vscale x 2 x i64> %op, i32 31, i32 16)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_s64,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecd_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecd_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svqdecd_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecd.nxv2i64(<vscale x 2 x i64> %op, i32 31, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd,_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecd_pat_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 8, i32 16)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecd_pat_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 8, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdecd_pat_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_pat_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecd.nxv2i64(<vscale x 2 x i64> %op, i32 8, i32 16)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd_pat,_s64,,)(op, SV_VL8, 16);
 }
 
+// CHECK-LABEL: @test_svqdecd_pat_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 9, i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecd_pat_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 9, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svqdecd_pat_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svqdecd_pat_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecd.nxv2i64(<vscale x 2 x i64> %op, i32 9, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecd_pat,_u64,,)(op, SV_VL16, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdech.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdech.c
index e29c38516350e..a406e38ca1d5f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdech.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdech.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,106 +14,197 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdech_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdech_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdech_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdech_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdech_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdech_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdech_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqdech_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdech_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdech.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdech_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdech.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdech_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdech_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdech.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdech_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdech.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdech_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdech.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdech_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdech_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdech.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdech_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdech.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdech_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdech.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdech_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdech_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdech.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdech_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 [[OP:%.*]], i32 10, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdech_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 [[OP:%.*]], i32 10, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdech_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdech_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdech.n32(i32 %op, i32 10, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech_pat,_n_s32,,)(op, SV_VL32, 16);
 }
 
+// CHECK-LABEL: @test_svqdech_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdech.n64(i64 [[OP:%.*]], i32 11, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdech_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdech.n64(i64 [[OP:%.*]], i32 11, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdech_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdech_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdech.n64(i64 %op, i32 11, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech_pat,_n_s64,,)(op, SV_VL64, 1);
 }
 
+// CHECK-LABEL: @test_svqdech_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdech.n32(i32 [[OP:%.*]], i32 12, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdech_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdech.n32(i32 [[OP:%.*]], i32 12, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdech_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdech_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdech.n32(i32 %op, i32 12, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech_pat,_n_u32,,)(op, SV_VL128, 16);
 }
 
+// CHECK-LABEL: @test_svqdech_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdech.n64(i64 [[OP:%.*]], i32 13, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdech_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdech.n64(i64 [[OP:%.*]], i32 13, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdech_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdech_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdech.n64(i64 %op, i32 13, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech_pat,_n_u64,,)(op, SV_VL256, 1);
 }
 
+// CHECK-LABEL: @test_svqdech_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdech_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdech_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svqdech_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdech.nxv8i16(<vscale x 8 x i16> %op, i32 31, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_s16,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdech_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdech_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqdech_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svqdech_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdech.nxv8i16(<vscale x 8 x i16> %op, i32 31, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech,_u16,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdech_pat_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 29, i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdech_pat_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 29, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdech_pat_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svqdech_pat_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdech.nxv8i16(<vscale x 8 x i16> %op, i32 29, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech_pat,_s16,,)(op, SV_MUL4, 16);
 }
 
+// CHECK-LABEL: @test_svqdech_pat_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 30, i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdech_pat_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdech.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 30, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqdech_pat_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svqdech_pat_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdech.nxv8i16(<vscale x 8 x i16> %op, i32 30, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdech_pat,_u16,,)(op, SV_MUL3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecp.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecp.c
index e6417e41eb2d8..e164504fddee8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecp.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecp.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,196 +14,368 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdecp_n_s32_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdecp_n_s32_b8iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecp_n_s32_b8(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s32_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv16i1(i32 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s32,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s32_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_s32_b16iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svqdecp_n_s32_b16(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s32_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv8i1(i32 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s32,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s32_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_s32_b32iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svqdecp_n_s32_b32(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s32_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv4i1(i32 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s32,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s32_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_s32_b64iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svqdecp_n_s32_b64(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s32_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecp.n32.nxv2i1(i32 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s32,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s64_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdecp_n_s64_b8lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecp_n_s64_b8(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s64_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv16i1(i64 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s64,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s64_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_s64_b16lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svqdecp_n_s64_b16(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s64_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv8i1(i64 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s64,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s64_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_s64_b32lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svqdecp_n_s64_b32(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s64_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv4i1(i64 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s64,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_s64_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_s64_b64lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svqdecp_n_s64_b64(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_s64_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecp.n64.nxv2i1(i64 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_s64,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u32_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdecp_n_u32_b8ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecp_n_u32_b8(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u32_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv16i1(i32 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u32,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u32_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_u32_b16ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svqdecp_n_u32_b16(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u32_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv8i1(i32 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u32,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u32_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_u32_b32ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svqdecp_n_u32_b32(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u32_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv4i1(i32 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u32,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u32_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_u32_b64ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svqdecp_n_u32_b64(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u32_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecp.n32.nxv2i1(i32 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u32,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u64_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdecp_n_u64_b8mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecp_n_u64_b8(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u64_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv16i1(i64 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u64,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u64_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_u64_b16mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svqdecp_n_u64_b16(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u64_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv8i1(i64 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u64,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u64_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_u64_b32mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svqdecp_n_u64_b32(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u64_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv4i1(i64 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u64,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_n_u64_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecp_n_u64_b64mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svqdecp_n_u64_b64(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_n_u64_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecp.n64.nxv2i1(i64 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_n_u64,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdecp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecp_s16u11__SVInt16_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdecp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdecp_s16(svint16_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdecp.nxv8i16(<vscale x 8 x i16> %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_s16,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecp_s32u11__SVInt32_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdecp_s32(svint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecp.nxv4i32(<vscale x 4 x i32> %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_s32,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecp_s64u11__SVInt64_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdecp_s64(svint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdecp.nxv2i64(<vscale x 2 x i64> %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_s64,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdecp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecp_u16u12__SVUint16_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdecp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqdecp_u16(svuint16_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqdecp.nxv8i16(<vscale x 8 x i16> %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_u16,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecp_u32u12__SVUint32_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqdecp_u32(svuint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecp.nxv4i32(<vscale x 4 x i32> %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_u32,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqdecp_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecp_u64u12__SVUint64_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqdecp_u64(svuint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqdecp_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqdecp.nxv2i64(<vscale x 2 x i64> %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecp,_u64,,)(op, pg);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecw.c
index 1208459f17d78..b3f0303bbf3c1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qdecw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,106 +14,197 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdecw_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecw_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecw_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecw_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecw_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecw_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecw_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecw_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecw_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdecw_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecw.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecw_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecw_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecw_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecw.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecw_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdecw_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecw_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdecw_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecw.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecw_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecw_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqdecw_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqdecw.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw_pat,_n_s32,,)(op, SV_ALL, 16);
 }
 
+// CHECK-LABEL: @test_svqdecw_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecw.n64(i64 [[OP:%.*]], i32 0, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecw_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqdecw.n64(i64 [[OP:%.*]], i32 0, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqdecw_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqdecw_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqdecw.n64(i64 %op, i32 0, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw_pat,_n_s64,,)(op, SV_POW2, 1);
 }
 
+// CHECK-LABEL: @test_svqdecw_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecw.n32(i32 [[OP:%.*]], i32 1, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecw_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqdecw.n32(i32 [[OP:%.*]], i32 1, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqdecw_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqdecw.n32(i32 %op, i32 1, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw_pat,_n_u32,,)(op, SV_VL1, 16);
 }
 
+// CHECK-LABEL: @test_svqdecw_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecw.n64(i64 [[OP:%.*]], i32 2, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdecw_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqdecw.n64(i64 [[OP:%.*]], i32 2, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqdecw_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqdecw_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqdecw.n64(i64 %op, i32 2, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw_pat,_n_u64,,)(op, SV_VL2, 1);
 }
 
+// CHECK-LABEL: @test_svqdecw_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecw_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdecw_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecw.nxv4i32(<vscale x 4 x i32> %op, i32 31, i32 16)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqdecw_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdecw_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqdecw_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecw.nxv4i32(<vscale x 4 x i32> %op, i32 31, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw,_u32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqdecw_pat_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 3, i32 16)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecw_pat_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 3, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdecw_pat_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_pat_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdecw.nxv4i32(<vscale x 4 x i32> %op, i32 3, i32 16)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw_pat,_s32,,)(op, SV_VL3, 16);
 }
 
+// CHECK-LABEL: @test_svqdecw_pat_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 4, i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdecw_pat_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 4, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqdecw_pat_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svqdecw_pat_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqdecw.nxv4i32(<vscale x 4 x i32> %op, i32 4, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqdecw_pat,_u32,,)(op, SV_VL4, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincb.c
index 3a18911029216..9e24a54665c82 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,74 +14,137 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqincb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincb_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincb_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqincb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincb_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincb_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqincb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincb_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincb_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqincb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincb.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincb_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincb.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincb_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqincb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincb.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincb_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincb.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincb_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqincb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincb.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincb_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 [[OP:%.*]], i32 5, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincb_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 [[OP:%.*]], i32 5, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincb_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqincb_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincb.n32(i32 %op, i32 5, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb_pat,_n_s32,,)(op, SV_VL5, 16);
 }
 
+// CHECK-LABEL: @test_svqincb_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincb.n64(i64 [[OP:%.*]], i32 6, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincb_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincb.n64(i64 [[OP:%.*]], i32 6, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincb_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqincb_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincb.n64(i64 %op, i32 6, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb_pat,_n_s64,,)(op, SV_VL6, 1);
 }
 
+// CHECK-LABEL: @test_svqincb_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincb.n32(i32 [[OP:%.*]], i32 7, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincb_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincb.n32(i32 [[OP:%.*]], i32 7, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincb_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqincb_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincb.n32(i32 %op, i32 7, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb_pat,_n_u32,,)(op, SV_VL7, 16);
 }
 
+// CHECK-LABEL: @test_svqincb_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincb.n64(i64 [[OP:%.*]], i32 8, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincb_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincb.n64(i64 [[OP:%.*]], i32 8, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincb_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqincb_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincb.n64(i64 %op, i32 8, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincb_pat,_n_u64,,)(op, SV_VL8, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincd.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincd.c
index 17f29bae301e6..c46818ce1ff7c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincd.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincd.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,106 +14,197 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqincd_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincd_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincd_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqincd_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincd_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincd_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincd_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqincd_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincd_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincd_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincd_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqincd_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincd.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincd_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincd_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincd.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincd_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqincd_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincd.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincd_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincd_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincd.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincd_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqincd_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincd.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincd_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 [[OP:%.*]], i32 9, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincd_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 [[OP:%.*]], i32 9, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincd_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqincd_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincd.n32(i32 %op, i32 9, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd_pat,_n_s32,,)(op, SV_VL16, 16);
 }
 
+// CHECK-LABEL: @test_svqincd_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincd.n64(i64 [[OP:%.*]], i32 10, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincd_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincd.n64(i64 [[OP:%.*]], i32 10, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincd_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqincd_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincd.n64(i64 %op, i32 10, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd_pat,_n_s64,,)(op, SV_VL32, 1);
 }
 
+// CHECK-LABEL: @test_svqincd_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincd.n32(i32 [[OP:%.*]], i32 11, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincd_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincd.n32(i32 [[OP:%.*]], i32 11, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincd_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqincd_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincd.n32(i32 %op, i32 11, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd_pat,_n_u32,,)(op, SV_VL64, 16);
 }
 
+// CHECK-LABEL: @test_svqincd_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincd.n64(i64 [[OP:%.*]], i32 12, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincd_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincd.n64(i64 [[OP:%.*]], i32 12, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincd_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqincd_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincd.n64(i64 %op, i32 12, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd_pat,_n_u64,,)(op, SV_VL128, 1);
 }
 
+// CHECK-LABEL: @test_svqincd_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincd_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqincd_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svqincd_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincd.nxv2i64(<vscale x 2 x i64> %op, i32 31, i32 16)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_s64,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincd_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincd_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svqincd_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svqincd_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincd.nxv2i64(<vscale x 2 x i64> %op, i32 31, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd,_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincd_pat_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 13, i32 16)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincd_pat_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 13, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqincd_pat_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svqincd_pat_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincd.nxv2i64(<vscale x 2 x i64> %op, i32 13, i32 16)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd_pat,_s64,,)(op, SV_VL256, 16);
 }
 
+// CHECK-LABEL: @test_svqincd_pat_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 29, i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincd_pat_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincd.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], i32 29, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svqincd_pat_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svqincd_pat_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincd.nxv2i64(<vscale x 2 x i64> %op, i32 29, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincd_pat,_u64,,)(op, SV_MUL4, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qinch.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qinch.c
index b5ccf09237c2d..c1ab033bb9ec9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qinch.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qinch.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,106 +14,197 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqinch_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqinch_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqinch_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqinch_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqinch_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqinch_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqinch_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqinch_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqinch_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqinch.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqinch_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqinch.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqinch_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqinch_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqinch.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqinch_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqinch.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqinch_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqinch.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqinch_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqinch_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqinch.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqinch_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqinch.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqinch_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqinch.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqinch_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqinch_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqinch.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqinch_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 [[OP:%.*]], i32 30, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqinch_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 [[OP:%.*]], i32 30, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqinch_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqinch_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqinch.n32(i32 %op, i32 30, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch_pat,_n_s32,,)(op, SV_MUL3, 16);
 }
 
+// CHECK-LABEL: @test_svqinch_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqinch.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqinch_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqinch.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqinch_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqinch_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqinch.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch_pat,_n_s64,,)(op, SV_ALL, 1);
 }
 
+// CHECK-LABEL: @test_svqinch_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqinch.n32(i32 [[OP:%.*]], i32 0, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqinch_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqinch.n32(i32 [[OP:%.*]], i32 0, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqinch_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqinch_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqinch.n32(i32 %op, i32 0, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch_pat,_n_u32,,)(op, SV_POW2, 16);
 }
 
+// CHECK-LABEL: @test_svqinch_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqinch.n64(i64 [[OP:%.*]], i32 1, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqinch_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqinch.n64(i64 [[OP:%.*]], i32 1, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqinch_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqinch_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqinch.n64(i64 %op, i32 1, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch_pat,_n_u64,,)(op, SV_VL1, 1);
 }
 
+// CHECK-LABEL: @test_svqinch_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqinch_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqinch_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svqinch_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqinch.nxv8i16(<vscale x 8 x i16> %op, i32 31, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_s16,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqinch_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqinch_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqinch_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svqinch_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqinch.nxv8i16(<vscale x 8 x i16> %op, i32 31, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch,_u16,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqinch_pat_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 2, i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqinch_pat_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 2, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqinch_pat_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svqinch_pat_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqinch.nxv8i16(<vscale x 8 x i16> %op, i32 2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch_pat,_s16,,)(op, SV_VL2, 16);
 }
 
+// CHECK-LABEL: @test_svqinch_pat_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 3, i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqinch_pat_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqinch.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], i32 3, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqinch_pat_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svqinch_pat_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqinch.nxv8i16(<vscale x 8 x i16> %op, i32 3, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqinch_pat,_u16,,)(op, SV_VL3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincp.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincp.c
index 55bbd48f4a196..f6bcd75332adb 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincp.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincp.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,196 +14,368 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqincp_n_s32_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqincp_n_s32_b8iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincp_n_s32_b8(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s32_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv16i1(i32 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s32,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s32_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_s32_b16iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svqincp_n_s32_b16(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s32_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv8i1(i32 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s32,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s32_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_s32_b32iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svqincp_n_s32_b32(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s32_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv4i1(i32 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s32,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s32_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_s32_b64iu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 int32_t test_svqincp_n_s32_b64(int32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s32_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincp.n32.nxv2i1(i32 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s32,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s64_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqincp_n_s64_b8lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincp_n_s64_b8(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s64_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv16i1(i64 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s64,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s64_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_s64_b16lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svqincp_n_s64_b16(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s64_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv8i1(i64 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s64,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s64_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_s64_b32lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svqincp_n_s64_b32(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s64_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv4i1(i64 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s64,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_s64_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_s64_b64lu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 int64_t test_svqincp_n_s64_b64(int64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_s64_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincp.n64.nxv2i1(i64 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_s64,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u32_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqincp_n_u32_b8ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv16i1(i32 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincp_n_u32_b8(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u32_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv16i1(i32 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u32,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u32_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_u32_b16ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv8i1(i32 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svqincp_n_u32_b16(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u32_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv8i1(i32 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u32,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u32_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_u32_b32ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv4i1(i32 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svqincp_n_u32_b32(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u32_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv4i1(i32 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u32,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u32_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i32 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_u32_b64ju10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv2i1(i32 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i32 [[TMP1]]
+//
 uint32_t test_svqincp_n_u32_b64(uint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u32_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincp.n32.nxv2i1(i32 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u32,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u64_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqincp_n_u64_b8mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv16i1(i64 [[OP:%.*]], <vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincp_n_u64_b8(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u64_b8
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv16i1(i64 %op, <vscale x 16 x i1> %pg)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u64,_b8,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u64_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_u64_b16mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv8i1(i64 [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svqincp_n_u64_b16(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u64_b16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv8i1(i64 %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u64,_b16,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u64_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_u64_b32mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv4i1(i64 [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svqincp_n_u64_b32(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u64_b32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv4i1(i64 %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u64,_b32,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_n_u64_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret i64 [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincp_n_u64_b64mu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv2i1(i64 [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret i64 [[TMP1]]
+//
 uint64_t test_svqincp_n_u64_b64(uint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_n_u64_b64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincp.n64.nxv2i1(i64 %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_n_u64,_b64,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqincp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincp_s16u11__SVInt16_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqincp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqincp_s16(svint16_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqincp.nxv8i16(<vscale x 8 x i16> %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_s16,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincp_s32u11__SVInt32_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqincp_s32(svint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincp.nxv4i32(<vscale x 4 x i32> %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_s32,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincp_s64u11__SVInt64_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqincp_s64(svint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqincp.nxv2i64(<vscale x 2 x i64> %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_s64,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqincp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincp_u16u12__SVUint16_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqincp.nxv8i16(<vscale x 8 x i16> [[OP:%.*]], <vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqincp_u16(svuint16_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqincp.nxv8i16(<vscale x 8 x i16> %op, <vscale x 8 x i1> %[[PG]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_u16,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincp_u32u12__SVUint32_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincp.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], <vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqincp_u32(svuint32_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincp.nxv4i32(<vscale x 4 x i32> %op, <vscale x 4 x i1> %[[PG]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_u32,,)(op, pg);
 }
 
+// CHECK-LABEL: @test_svqincp_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincp_u64u12__SVUint64_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincp.nxv2i64(<vscale x 2 x i64> [[OP:%.*]], <vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqincp_u64(svuint64_t op, svbool_t pg)
 {
-  // CHECK-LABEL: test_svqincp_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqincp.nxv2i64(<vscale x 2 x i64> %op, <vscale x 2 x i1> %[[PG]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincp,_u64,,)(op, pg);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincw.c
index 712bdfa521429..990ea55a55569 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qincw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,106 +14,197 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqincw_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincw_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincw_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqincw_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 %op, i32 31, i32 1)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_n_s32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincw_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincw_n_s32_1i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincw_n_s32_1(int32_t op)
 {
-  // CHECK-LABEL: test_svqincw_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_n_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincw_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincw_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincw_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqincw_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincw.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_n_s64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincw_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincw_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincw.n32(i32 [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincw_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqincw_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincw.n32(i32 %op, i32 31, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_n_u32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincw_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqincw_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincw.n64(i64 [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincw_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqincw_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincw.n64(i64 %op, i32 31, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_n_u64,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincw_pat_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 [[OP:%.*]], i32 4, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincw_pat_n_s32i(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 [[OP:%.*]], i32 4, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 int32_t test_svqincw_pat_n_s32(int32_t op)
 {
-  // CHECK-LABEL: test_svqincw_pat_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.sqincw.n32(i32 %op, i32 4, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw_pat,_n_s32,,)(op, SV_VL4, 16);
 }
 
+// CHECK-LABEL: @test_svqincw_pat_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincw.n64(i64 [[OP:%.*]], i32 5, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincw_pat_n_s64l(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.sqincw.n64(i64 [[OP:%.*]], i32 5, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 int64_t test_svqincw_pat_n_s64(int64_t op)
 {
-  // CHECK-LABEL: test_svqincw_pat_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.sqincw.n64(i64 %op, i32 5, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw_pat,_n_s64,,)(op, SV_VL5, 1);
 }
 
+// CHECK-LABEL: @test_svqincw_pat_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincw.n32(i32 [[OP:%.*]], i32 6, i32 16)
+// CHECK-NEXT:    ret i32 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincw_pat_n_u32j(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i32 @llvm.aarch64.sve.uqincw.n32(i32 [[OP:%.*]], i32 6, i32 16)
+// CPP-CHECK-NEXT:    ret i32 [[TMP0]]
+//
 uint32_t test_svqincw_pat_n_u32(uint32_t op)
 {
-  // CHECK-LABEL: test_svqincw_pat_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call i32 @llvm.aarch64.sve.uqincw.n32(i32 %op, i32 6, i32 16)
-  // CHECK: ret i32 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw_pat,_n_u32,,)(op, SV_VL6, 16);
 }
 
+// CHECK-LABEL: @test_svqincw_pat_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincw.n64(i64 [[OP:%.*]], i32 7, i32 1)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqincw_pat_n_u64m(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call i64 @llvm.aarch64.sve.uqincw.n64(i64 [[OP:%.*]], i32 7, i32 1)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
 uint64_t test_svqincw_pat_n_u64(uint64_t op)
 {
-  // CHECK-LABEL: test_svqincw_pat_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call i64 @llvm.aarch64.sve.uqincw.n64(i64 %op, i32 7, i32 1)
-  // CHECK: ret i64 %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw_pat,_n_u64,,)(op, SV_VL7, 1);
 }
 
+// CHECK-LABEL: @test_svqincw_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 16)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincw_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqincw_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svqincw_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincw.nxv4i32(<vscale x 4 x i32> %op, i32 31, i32 16)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_s32,,)(op, 16);
 }
 
+// CHECK-LABEL: @test_svqincw_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqincw_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 31, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqincw_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svqincw_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincw.nxv4i32(<vscale x 4 x i32> %op, i32 31, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw,_u32,,)(op, 1);
 }
 
+// CHECK-LABEL: @test_svqincw_pat_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 8, i32 16)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincw_pat_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 8, i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqincw_pat_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svqincw_pat_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqincw.nxv4i32(<vscale x 4 x i32> %op, i32 8, i32 16)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw_pat,_s32,,)(op, SV_VL8, 16);
 }
 
+// CHECK-LABEL: @test_svqincw_pat_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 9, i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqincw_pat_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincw.nxv4i32(<vscale x 4 x i32> [[OP:%.*]], i32 9, i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqincw_pat_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svqincw_pat_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqincw.nxv4i32(<vscale x 4 x i32> %op, i32 9, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqincw_pat,_u32,,)(op, SV_VL16, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qsub.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qsub.c
index 3a59667f3c2fd..8ec7f0f377dd2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qsub.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_qsub.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,138 +14,258 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqsub_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svqsub_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqsub_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqsub_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqsub_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqsub_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqsub_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqsub_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqsub_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svqsub_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqsub_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqsub_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqsub_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqsub_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqsub_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqsub_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svqsub_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_n_s8u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsub_n_s8(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqsub_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqsub_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqsub_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsub_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqsub_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqsub_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqsub_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svqsub,_n_u64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rbit.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rbit.c
index bbb6f9765be52..46dfe8158b324 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rbit.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rbit.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,212 +14,398 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrbit_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrbit_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrbit_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svrbit_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrbit_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svrbit_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrbit_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrbit_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrbit_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrbit_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrbit_u8_zu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrbit_u8_z(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svrbit_u8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrbit_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svrbit_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrbit_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrbit_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrbit_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrbit_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrbit_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrbit_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svrbit_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrbit_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svrbit_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrbit_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrbit_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrbit_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrbit_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrbit_u8_mu11__SVUint8_tu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrbit_u8_m(svuint8_t inactive, svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svrbit_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrbit_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svrbit_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrbit_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrbit_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrbit_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrbit_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrbit_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrbit_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svrbit_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrbit_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svrbit_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrbit_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrbit_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrbit_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrbit_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrbit_u8_xu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrbit_u8_x(svbool_t pg, svuint8_t op)
 {
-  // CHECK-LABEL: test_svrbit_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rbit.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrbit_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svrbit_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rbit.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrbit_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrbit_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rbit.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrbit_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrbit_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrbit_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrbit_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rbit.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrbit,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rdffr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rdffr.c
index a85ac7bb5cef0..00cbc1f221aa3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rdffr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rdffr.c
@@ -1,22 +1,38 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svrdffr(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rdffr.z(<vscale x 16 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z12test_svrdffrv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rdffr.z(<vscale x 16 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svrdffr()
 {
-  // CHECK-LABEL: test_svrdffr
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rdffr.z(<vscale x 16 x i1>
-  // CHECK-NOT: rdffr
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svrdffr();
 }
 
+// CHECK-LABEL: @test_svrdffr_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rdffr.z(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrdffr_zu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rdffr.z(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svrdffr_z(svbool_t pg)
 {
-  // CHECK-LABEL: test_svrdffr_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rdffr.z(<vscale x 16 x i1> %pg)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svrdffr_z(pg);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpe.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpe.c
index 202ae69ac3c63..d38879b1f1946 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpe.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpe.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrecpe_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpe.x.nxv8f16(<vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrecpe_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpe.x.nxv8f16(<vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svrecpe_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrecpe_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpe.x.nxv8f16(<vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpe,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svrecpe_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpe.x.nxv4f32(<vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrecpe_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpe.x.nxv4f32(<vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svrecpe_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrecpe_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpe.x.nxv4f32(<vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpe,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svrecpe_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpe.x.nxv2f64(<vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrecpe_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpe.x.nxv2f64(<vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svrecpe_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrecpe_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpe.x.nxv2f64(<vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpe,_f64,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recps.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recps.c
index 93415ea0a039a..4e7effa982f61 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recps.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recps.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrecps_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecps.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrecps_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecps.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svrecps_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svrecps_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecps.x.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecps,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrecps_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecps.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrecps_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecps.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svrecps_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svrecps_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecps.x.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecps,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrecps_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecps.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrecps_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecps.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svrecps_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svrecps_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecps.x.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecps,_f64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpx.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpx.c
index 0a25ca42e2123..60ca8534a1d02 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpx.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_recpx.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrecpx_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrecpx_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrecpx_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrecpx_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrecpx_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrecpx_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrecpx_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrecpx_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frecpx.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrecpx_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frecpx.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpx_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpx_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrecpx_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrecpx_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frecpx.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrecpx,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret-bfloat.c
index da1aa4b7b5408..3d903febb47fe 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -15,161 +16,321 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svreinterpret_s8_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s8_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_s8_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_s16_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_s16_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_s32_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_s32_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32, _bf16, , )(op);
 }
+// CHECK-LABEL: @test_svreinterpret_s64_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_s64_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_s64_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u8_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_u8_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_u16_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_u16_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_u32_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_u32_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_u64_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_u64_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_bf16_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_s8(svint8_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _s8, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_s16(svint16_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _s16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_s32(svint32_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _s32, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_s64(svint64_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _s64, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_bf16_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_u8(svuint8_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _u8, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_u16(svuint16_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _u16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_u32(svuint32_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _u32, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_u64(svuint64_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _u64, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z28test_svreinterpret_bf16_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[OP:%.*]]
+//
 svbfloat16_t test_svreinterpret_bf16_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_bf16
-  // CHECK: ret <vscale x 8 x bfloat> %op
   return SVE_ACLE_FUNC(svreinterpret_bf16, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_f16(svfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _f16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_f32(svfloat32_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _f32, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_bf16_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_bf16_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x bfloat>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svreinterpret_bf16_f64(svfloat64_t op) {
-  // CHECK-LABEL: test_svreinterpret_bf16_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 8 x bfloat>
-  // CHECK: ret <vscale x 8 x bfloat> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_bf16, _f64, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_f32_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_f32_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_f16_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_f16_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16, _bf16, , )(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z27test_svreinterpret_f64_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x bfloat> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_bf16(svbfloat16_t op) {
-  // CHECK-LABEL: test_svreinterpret_f64_bf16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x bfloat> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64, _bf16, , )(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret.c
index 11c376092a625..0a6847943d269 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_reinterpret.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,951 +15,1786 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svreinterpret_s8_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svreinterpret_s8_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
 svint8_t test_svreinterpret_s8_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_s8
-  // CHECK: ret <vscale x 16 x i8> %op
   return SVE_ACLE_FUNC(svreinterpret_s8,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svreinterpret_s8_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
 svint8_t test_svreinterpret_s8_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_u8
-  // CHECK: ret <vscale x 16 x i8> %op
   return SVE_ACLE_FUNC(svreinterpret_s8,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s8_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s8_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svreinterpret_s8_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s8_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s8,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s16_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
 svint16_t test_svreinterpret_s16_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_s16
-  // CHECK: ret <vscale x 8 x i16> %op
   return SVE_ACLE_FUNC(svreinterpret_s16,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s16_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
 svint16_t test_svreinterpret_s16_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_u16
-  // CHECK: ret <vscale x 8 x i16> %op
   return SVE_ACLE_FUNC(svreinterpret_s16,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s16_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s16_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svreinterpret_s16_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s16_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s16,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s32_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
 svint32_t test_svreinterpret_s32_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_s32
-  // CHECK: ret <vscale x 4 x i32> %op
   return SVE_ACLE_FUNC(svreinterpret_s32,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s32_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
 svint32_t test_svreinterpret_s32_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_u32
-  // CHECK: ret <vscale x 4 x i32> %op
   return SVE_ACLE_FUNC(svreinterpret_s32,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s32_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s32_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svreinterpret_s32_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s32_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s32,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s64_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
 svint64_t test_svreinterpret_s64_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_s64
-  // CHECK: ret <vscale x 2 x i64> %op
   return SVE_ACLE_FUNC(svreinterpret_s64,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_s64_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
 svint64_t test_svreinterpret_s64_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_u64
-  // CHECK: ret <vscale x 2 x i64> %op
   return SVE_ACLE_FUNC(svreinterpret_s64,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_s64_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_s64_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svreinterpret_s64_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_s64_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_s64,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svreinterpret_u8_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
 svuint8_t test_svreinterpret_u8_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_s8
-  // CHECK: ret <vscale x 16 x i8> %op
   return SVE_ACLE_FUNC(svreinterpret_u8,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svreinterpret_u8_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[OP:%.*]]
+//
 svuint8_t test_svreinterpret_u8_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_u8
-  // CHECK: ret <vscale x 16 x i8> %op
   return SVE_ACLE_FUNC(svreinterpret_u8,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u8_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 16 x i8>
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u8_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 16 x i8>
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svreinterpret_u8_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u8_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 16 x i8>
-  // CHECK: ret <vscale x 16 x i8> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u8,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u16_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
 svuint16_t test_svreinterpret_u16_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_s16
-  // CHECK: ret <vscale x 8 x i16> %op
   return SVE_ACLE_FUNC(svreinterpret_u16,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u16_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[OP:%.*]]
+//
 svuint16_t test_svreinterpret_u16_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_u16
-  // CHECK: ret <vscale x 8 x i16> %op
   return SVE_ACLE_FUNC(svreinterpret_u16,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u16_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u16_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svreinterpret_u16_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u16_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u16,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u32_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
 svuint32_t test_svreinterpret_u32_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_s32
-  // CHECK: ret <vscale x 4 x i32> %op
   return SVE_ACLE_FUNC(svreinterpret_u32,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u32_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[OP:%.*]]
+//
 svuint32_t test_svreinterpret_u32_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_u32
-  // CHECK: ret <vscale x 4 x i32> %op
   return SVE_ACLE_FUNC(svreinterpret_u32,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u32_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u32_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svreinterpret_u32_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u32_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u32,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u64_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
 svuint64_t test_svreinterpret_u64_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_s64
-  // CHECK: ret <vscale x 2 x i64> %op
   return SVE_ACLE_FUNC(svreinterpret_u64,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_u64_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[OP:%.*]]
+//
 svuint64_t test_svreinterpret_u64_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_u64
-  // CHECK: ret <vscale x 2 x i64> %op
   return SVE_ACLE_FUNC(svreinterpret_u64,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_u64_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_u64_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svreinterpret_u64_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_u64_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_u64,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_f16_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_s32,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_s64,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f16_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_f16_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_u8,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f16_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_u16,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_u32,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x half> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[OP:%.*]]
+//
 svfloat16_t test_svreinterpret_f16_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_f16
-  // CHECK: ret <vscale x 8 x half> %op
   return SVE_ACLE_FUNC(svreinterpret_f16,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f16_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x half>
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f16_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 8 x half>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svreinterpret_f16_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f16_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 8 x half>
-  // CHECK: ret <vscale x 8 x half> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f16,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_f32_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_f32_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f32_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x float> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[OP:%.*]]
+//
 svfloat32_t test_svreinterpret_f32_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_f32
-  // CHECK: ret <vscale x 4 x float> %op
   return SVE_ACLE_FUNC(svreinterpret_f32,_f32,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f32_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 4 x float>
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f32_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x double> [[OP:%.*]] to <vscale x 4 x float>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svreinterpret_f32_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f32_f64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x double> %op to <vscale x 4 x float>
-  // CHECK: ret <vscale x 4 x float> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f32,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_f64_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_s8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_s16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_s32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_s64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svreinterpret_f64_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 16 x i8> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_u8
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 16 x i8> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x i16> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_u16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x i16> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x i32> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_u32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x i32> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 2 x i64> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_u64
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 2 x i64> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_u64,,)(op);
 }
 
+//
+// CHECK-LABEL: @test_svreinterpret_f64_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 8 x half> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_f16
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 8 x half> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 2 x double>
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast <vscale x 4 x float> [[OP:%.*]] to <vscale x 2 x double>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svreinterpret_f64_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_f32
-  // CHECK: %[[CAST:.*]] = bitcast <vscale x 4 x float> %op to <vscale x 2 x double>
-  // CHECK: ret <vscale x 2 x double> %[[CAST]]
   return SVE_ACLE_FUNC(svreinterpret_f64,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svreinterpret_f64_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x double> [[OP:%.*]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svreinterpret_f64_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[OP:%.*]]
+//
 svfloat64_t test_svreinterpret_f64_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svreinterpret_f64_f64
-  // CHECK: ret <vscale x 2 x double> %op
   return SVE_ACLE_FUNC(svreinterpret_f64,_f64,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev-bfloat.c
index 22b9dd7621e75..d8543546decf0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrev_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.rev.nxv8bf16(<vscale x 8 x bfloat> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svrev_bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.rev.nxv8bf16(<vscale x 8 x bfloat> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svrev_bf16(svbfloat16_t op)
 {
-  // CHECK-LABEL: test_svrev_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.rev.nxv8bf16(<vscale x 8 x bfloat> %op)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svrev_bf16'}}
   return SVE_ACLE_FUNC(svrev,_bf16,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev.c
index c532cd0a015df..04dd6da4332c2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rev.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,128 +14,239 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrev_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rev.nxv16i8(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svrev_s8u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rev.nxv16i8(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrev_s8(svint8_t op)
 {
-  // CHECK-LABEL: test_svrev_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rev.nxv16i8(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_s8,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rev.nxv8i16(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rev.nxv8i16(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrev_s16(svint16_t op)
 {
-  // CHECK-LABEL: test_svrev_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rev.nxv8i16(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rev.nxv4i32(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rev.nxv4i32(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrev_s32(svint32_t op)
 {
-  // CHECK-LABEL: test_svrev_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rev.nxv4i32(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rev.nxv2i64(<vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rev.nxv2i64(<vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svrev_s64(svint64_t op)
 {
-  // CHECK-LABEL: test_svrev_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rev.nxv2i64(<vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rev.nxv16i8(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svrev_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rev.nxv16i8(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrev_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svrev_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rev.nxv16i8(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_u8,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rev.nxv8i16(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rev.nxv8i16(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrev_u16(svuint16_t op)
 {
-  // CHECK-LABEL: test_svrev_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rev.nxv8i16(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rev.nxv4i32(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rev.nxv4i32(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrev_u32(svuint32_t op)
 {
-  // CHECK-LABEL: test_svrev_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rev.nxv4i32(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rev.nxv2i64(<vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rev.nxv2i64(<vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svrev_u64(svuint64_t op)
 {
-  // CHECK-LABEL: test_svrev_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rev.nxv2i64(<vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.rev.nxv8f16(<vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.rev.nxv8f16(<vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svrev_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrev_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.rev.nxv8f16(<vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.rev.nxv4f32(<vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.rev.nxv4f32(<vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svrev_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrev_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.rev.nxv4f32(<vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.rev.nxv2f64(<vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.rev.nxv2f64(<vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svrev_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrev_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.rev.nxv2f64(<vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrev,_f64,,)(op);
 }
 
+// CHECK-LABEL: @test_svrev_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rev.nxv16i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svrev_b8u10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rev.nxv16i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svrev_b8(svbool_t op)
 {
-  // CHECK-LABEL: test_svrev_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.rev.nxv16i1(<vscale x 16 x i1> %op)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svrev_b8(op);
 }
 
+// CHECK-LABEL: @test_svrev_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.rev.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_b16u10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.rev.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svrev_b16(svbool_t op)
 {
-  // CHECK-LABEL: test_svrev_b16
-  // CHECK: %[[OP:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.rev.nxv8i1(<vscale x 8 x i1> %[[OP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svrev_b16(op);
 }
 
+// CHECK-LABEL: @test_svrev_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.rev.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_b32u10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.rev.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svrev_b32(svbool_t op)
 {
-  // CHECK-LABEL: test_svrev_b32
-  // CHECK: %[[OP:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.rev.nxv4i1(<vscale x 4 x i1> %[[OP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svrev_b32(op);
 }
 
+// CHECK-LABEL: @test_svrev_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.rev.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svrev_b64u10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.rev.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svrev_b64(svbool_t op)
 {
-  // CHECK-LABEL: test_svrev_b64
-  // CHECK: %[[OP:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.rev.nxv2i1(<vscale x 2 x i1> %[[OP]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svrev_b64(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revb.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revb.c
index 59b65015660bf..eef7681648ef8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revb.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revb.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,164 +14,308 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrevb_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrevb_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svrevb_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrevb_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrevb_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevb_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevb_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrevb_u16_z(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svrevb_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrevb_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrevb_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevb_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevb_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrevb_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svrevb_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrevb_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrevb_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevb_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevb_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u16_mu12__SVUint16_tu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrevb_u16_m(svuint16_t inactive, svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svrevb_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrevb_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrevb_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevb_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevb_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrevb_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svrevb_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrevb_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrevb_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevb_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevb_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrevb_u16_x(svbool_t pg, svuint16_t op)
 {
-  // CHECK-LABEL: test_svrevb_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.revb.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrevb_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrevb_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revb.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevb_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevb_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevb_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevb_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revb.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevb,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revh.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revh.c
index 3cada1f083003..5bb7a046d5d28 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revh.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revh.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,110 +14,206 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrevh_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrevh_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrevh_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevh_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevh_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrevh_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrevh_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevh_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevh_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrevh_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrevh_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevh_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevh_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrevh_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrevh_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevh_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevh_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrevh_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svrevh_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevh_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevh_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrevh_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrevh_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.revh.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_u32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevh_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevh_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevh_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevh_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revh.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevh,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revw.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revw.c
index 8a131a85657b5..852a0ea0a8098 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revw.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_revw.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,56 +14,104 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrevw_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevw_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevw_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevw_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevw,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevw_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevw_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevw_u64_z(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevw_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevw,_u64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevw_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevw_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevw_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevw_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevw,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevw_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevw_u64_mu12__SVUint64_tu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevw_u64_m(svuint64_t inactive, svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevw_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevw,_u64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrevw_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevw_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrevw_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svrevw_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevw,_s64,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrevw_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrevw_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrevw_u64_x(svbool_t pg, svuint64_t op)
 {
-  // CHECK-LABEL: test_svrevw_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.revw.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrevw,_u64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinta.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinta.c
index 377b15f91f201..6da6c2b4f15c1 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinta.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinta.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrinta_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrinta_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrinta_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrinta_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrinta_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrinta_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrinta_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrinta_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrinta_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrinta_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrinta_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrinta_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrinta_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrinta_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrinta_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinta.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrinta_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrinta_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinta.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinta_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinta_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrinta_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrinta_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinta.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinta,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinti.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinti.c
index 5b587fa0526f9..1a6bb1215804a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinti.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rinti.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrinti_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrinti_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrinti_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrinti_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrinti_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrinti_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrinti_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrinti_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrinti_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrinti_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrinti_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrinti_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrinti_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrinti_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrinti_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frinti.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrinti_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrinti_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frinti.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrinti_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrinti_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrinti_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrinti_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frinti.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrinti,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintm.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintm.c
index f1276ff4a0d82..27aac220cfa26 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintm.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintm.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrintm_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintm_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintm_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintm_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintm_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintm_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintm_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintm_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintm_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintm_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintm_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintm_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintm_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintm_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintm_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintm.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintm_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintm_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintm.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintm_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintm_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintm_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintm_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintm.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintm,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintn.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintn.c
index b519f8f7471c6..8caed46c58481 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintn.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintn.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrintn_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintn_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintn_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintn_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintn_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintn_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintn_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintn_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintn_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintn_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintn_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintn_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintn_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintn_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintn_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintn.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintn_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintn_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintn.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintn_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintn_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintn_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintn_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintn.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintn,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintp.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintp.c
index bda2c151f904d..e7962934442a6 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintp.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintp.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrintp_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintp_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintp_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintp_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintp_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintp_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintp_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintp_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintp_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintp_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintp_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintp_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintp_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintp_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintp_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintp.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintp_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintp_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintp.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintp_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintp_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintp_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintp_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintp.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintp,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintx.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintx.c
index 3cfdd099596d6..b3e3e6e9cec0a 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintx.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintx.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrintx_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintx_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintx_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintx_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintx_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintx_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintx_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintx_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintx_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintx_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintx_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintx_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintx_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintx_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintx_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintx.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintx_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintx_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintx.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintx_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintx_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintx_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintx_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintx.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintx,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintz.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintz.c
index bbf60c78ee8a2..143719d2a2f26 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintz.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rintz.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrintz_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintz_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintz_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintz_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintz_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintz_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintz_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintz_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintz_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintz_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintz_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintz_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintz_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svrintz_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrintz_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frintz.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svrintz_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrintz_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frintz.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrintz_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrintz_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svrintz_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrintz_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frintz.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrintz,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrte.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrte.c
index 3f4d6cb6289d7..e8cbb85c337bc 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrte.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrte.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrsqrte_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frsqrte.x.nxv8f16(<vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsqrte_f16u13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frsqrte.x.nxv8f16(<vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svrsqrte_f16(svfloat16_t op)
 {
-  // CHECK-LABEL: test_svrsqrte_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frsqrte.x.nxv8f16(<vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrsqrte,_f16,,)(op);
 }
 
+// CHECK-LABEL: @test_svrsqrte_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frsqrte.x.nxv4f32(<vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsqrte_f32u13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frsqrte.x.nxv4f32(<vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svrsqrte_f32(svfloat32_t op)
 {
-  // CHECK-LABEL: test_svrsqrte_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frsqrte.x.nxv4f32(<vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrsqrte,_f32,,)(op);
 }
 
+// CHECK-LABEL: @test_svrsqrte_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frsqrte.x.nxv2f64(<vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsqrte_f64u13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frsqrte.x.nxv2f64(<vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svrsqrte_f64(svfloat64_t op)
 {
-  // CHECK-LABEL: test_svrsqrte_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frsqrte.x.nxv2f64(<vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrsqrte,_f64,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrts.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrts.c
index 481b13acf9173..d378743d4aa40 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrts.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_rsqrts.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrsqrts_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frsqrts.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsqrts_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frsqrts.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svrsqrts_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svrsqrts_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.frsqrts.x.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrsqrts,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsqrts_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frsqrts.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsqrts_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frsqrts.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svrsqrts_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svrsqrts_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.frsqrts.x.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrsqrts,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsqrts_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frsqrts.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsqrts_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frsqrts.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svrsqrts_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svrsqrts_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.frsqrts.x.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svrsqrts,_f64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_scale.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_scale.c
index ef53940b9fa07..035adedd821ed 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_scale.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_scale.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,179 +14,338 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svscale_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f16_zu10__SVBool_tu13__SVFloat16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svscale_f16_z(svbool_t pg, svfloat16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svscale_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f32_zu10__SVBool_tu13__SVFloat32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svscale_f32_z(svbool_t pg, svfloat32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svscale_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f64_zu10__SVBool_tu13__SVFloat64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svscale_f64_z(svbool_t pg, svfloat64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svscale_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f16_mu10__SVBool_tu13__SVFloat16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svscale_f16_m(svbool_t pg, svfloat16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svscale_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f32_mu10__SVBool_tu13__SVFloat32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svscale_f32_m(svbool_t pg, svfloat32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svscale_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f64_mu10__SVBool_tu13__SVFloat64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svscale_f64_m(svbool_t pg, svfloat64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svscale_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f16_xu10__SVBool_tu13__SVFloat16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svscale_f16_x(svbool_t pg, svfloat16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svscale_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f32_xu10__SVBool_tu13__SVFloat32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svscale_f32_x(svbool_t pg, svfloat32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svscale_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svscale_f64_xu10__SVBool_tu13__SVFloat64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svscale_f64_x(svbool_t pg, svfloat64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svscale_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f16_zu10__SVBool_tu13__SVFloat16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svscale_n_f16_z(svbool_t pg, svfloat16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f32_zu10__SVBool_tu13__SVFloat32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svscale_n_f32_z(svbool_t pg, svfloat32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f64_zu10__SVBool_tu13__SVFloat64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svscale_n_f64_z(svbool_t pg, svfloat64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f16_mu10__SVBool_tu13__SVFloat16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svscale_n_f16_m(svbool_t pg, svfloat16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f32_mu10__SVBool_tu13__SVFloat32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svscale_n_f32_m(svbool_t pg, svfloat32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f64_mu10__SVBool_tu13__SVFloat64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svscale_n_f64_m(svbool_t pg, svfloat64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f16_xu10__SVBool_tu13__SVFloat16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svscale_n_f16_x(svbool_t pg, svfloat16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fscale.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f32_xu10__SVBool_tu13__SVFloat32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svscale_n_f32_x(svbool_t pg, svfloat32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fscale.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svscale_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svscale_n_f64_xu10__SVBool_tu13__SVFloat64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svscale_n_f64_x(svbool_t pg, svfloat64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svscale_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fscale.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svscale,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel-bfloat.c
index d95ff53acf00e..4b4cf34dc4a23 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel-bfloat.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,12 +14,20 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsel_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.sel.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsel_bf16u10__SVBool_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.sel.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svsel_bf16(svbool_t pg, svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsel_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.sel.nxv8bf16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svsel_bf16'}}
   return SVE_ACLE_FUNC(svsel,_bf16,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel.c
index 86a02d7265883..f252aaccf1bbd 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sel.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,107 +14,200 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsel_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svsel_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsel_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsel_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsel_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsel_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsel_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsel_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsel_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsel_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svsel_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsel_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsel_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsel_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsel_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsel_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsel_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsel_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsel_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsel_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsel_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsel_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsel_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svsel_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsel_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsel_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_f64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsel_b(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.sel.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z12test_svsel_bu10__SVBool_tu10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.sel.nxv16i1(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svsel_b(svbool_t pg, svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svsel_b
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.sel.nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsel,_b,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
index 0cbd71ed0029e..676f2c26bef4f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,20 +14,34 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svset2_bf16_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset2_bf16_014svbfloat16x2_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
 svbfloat16x2_t test_svset2_bf16_0(svbfloat16x2_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset2_bf16_0
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> %tuple, i32 0, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 16 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset2_bf16'}}
   return SVE_ACLE_FUNC(svset2,_bf16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset2_bf16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset2_bf16_114svbfloat16x2_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
 svbfloat16x2_t test_svset2_bf16_1(svbfloat16x2_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset2_bf16_1
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x bfloat> @llvm.aarch64.sve.tuple.set.nxv16bf16.nxv8bf16(<vscale x 16 x bfloat> %tuple, i32 1, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 16 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset2_bf16'}}
   return SVE_ACLE_FUNC(svset2,_bf16,,)(tuple, 1, x);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
index 175341b108d36..de81dedfba167 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set2.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,90 +13,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svset2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svset2_s810svint8x2_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
 svint8x2_t test_svset2_s8(svint8x2_t tuple, svint8_t x)
 {
-  // CHECK-LABEL: test_svset2_s8
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> %tuple, i32 1, <vscale x 16 x i8> %x)
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_s8,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_s1611svint16x2_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
 svint16x2_t test_svset2_s16(svint16x2_t tuple, svint16_t x)
 {
-  // CHECK-LABEL: test_svset2_s16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> %tuple, i32 0, <vscale x 8 x i16> %x)
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_s16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_s3211svint32x2_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
 svint32x2_t test_svset2_s32(svint32x2_t tuple, svint32_t x)
 {
-  // CHECK-LABEL: test_svset2_s32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> %tuple, i32 1, <vscale x 4 x i32> %x)
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_s32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_s6411svint64x2_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
 svint64x2_t test_svset2_s64(svint64x2_t tuple, svint64_t x)
 {
-  // CHECK-LABEL: test_svset2_s64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> %tuple, i32 0, <vscale x 2 x i64> %x)
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_s64,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svset2_u811svuint8x2_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
 svuint8x2_t test_svset2_u8(svuint8x2_t tuple, svuint8_t x)
 {
-  // CHECK-LABEL: test_svset2_u8
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x i8> @llvm.aarch64.sve.tuple.set.nxv32i8.nxv16i8(<vscale x 32 x i8> %tuple, i32 1, <vscale x 16 x i8> %x)
-  // CHECK-NEXT: ret <vscale x 32 x i8> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_u8,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_u1612svuint16x2_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
 svuint16x2_t test_svset2_u16(svuint16x2_t tuple, svuint16_t x)
 {
-  // CHECK-LABEL: test_svset2_u16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x i16> @llvm.aarch64.sve.tuple.set.nxv16i16.nxv8i16(<vscale x 16 x i16> %tuple, i32 0, <vscale x 8 x i16> %x)
-  // CHECK-NEXT: ret <vscale x 16 x i16> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_u16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_u3212svuint32x2_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
 svuint32x2_t test_svset2_u32(svuint32x2_t tuple, svuint32_t x)
 {
-  // CHECK-LABEL: test_svset2_u32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 8 x i32> @llvm.aarch64.sve.tuple.set.nxv8i32.nxv4i32(<vscale x 8 x i32> %tuple, i32 1, <vscale x 4 x i32> %x)
-  // CHECK-NEXT: ret <vscale x 8 x i32> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_u32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_u6412svuint64x2_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> [[TUPLE:%.*]], i32 0, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
 svuint64x2_t test_svset2_u64(svuint64x2_t tuple, svuint64_t x)
 {
-  // CHECK-LABEL: test_svset2_u64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 4 x i64> @llvm.aarch64.sve.tuple.set.nxv4i64.nxv2i64(<vscale x 4 x i64> %tuple, i32 0, <vscale x 2 x i64> %x)
-  // CHECK-NEXT: ret <vscale x 4 x i64> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_u64,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.tuple.set.nxv16f16.nxv8f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_f1613svfloat16x2_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.tuple.set.nxv16f16.nxv8f16(<vscale x 16 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x half> [[TMP0]]
+//
 svfloat16x2_t test_svset2_f16(svfloat16x2_t tuple, svfloat16_t x)
 {
-  // CHECK-LABEL: test_svset2_f16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x half> @llvm.aarch64.sve.tuple.set.nxv16f16.nxv8f16(<vscale x 16 x half> %tuple, i32 1, <vscale x 8 x half> %x)
-  // CHECK-NEXT: ret <vscale x 16 x half> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_f16,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.tuple.set.nxv8f32.nxv4f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 0, <vscale x 4 x float> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_f3213svfloat32x2_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.tuple.set.nxv8f32.nxv4f32(<vscale x 8 x float> [[TUPLE:%.*]], i32 0, <vscale x 4 x float> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
 svfloat32x2_t test_svset2_f32(svfloat32x2_t tuple, svfloat32_t x)
 {
-  // CHECK-LABEL: test_svset2_f32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 8 x float> @llvm.aarch64.sve.tuple.set.nxv8f32.nxv4f32(<vscale x 8 x float> %tuple, i32 0, <vscale x 4 x float> %x)
-  // CHECK-NEXT: ret <vscale x 8 x float> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_f32,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.tuple.set.nxv4f64.nxv2f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 1, <vscale x 2 x double> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset2_f6413svfloat64x2_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.tuple.set.nxv4f64.nxv2f64(<vscale x 4 x double> [[TUPLE:%.*]], i32 1, <vscale x 2 x double> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x double> [[TMP0]]
+//
 svfloat64x2_t test_svset2_f64(svfloat64x2_t tuple, svfloat64_t x)
 {
-  // CHECK-LABEL: test_svset2_f64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 4 x double> @llvm.aarch64.sve.tuple.set.nxv4f64.nxv2f64(<vscale x 4 x double> %tuple, i32 1, <vscale x 2 x double> %x)
-  // CHECK-NEXT: ret <vscale x 4 x double> %[[INSERT]]
   return SVE_ACLE_FUNC(svset2,_f64,,)(tuple, 1, x);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
index da1f81a6600e3..a2693c6979517 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -14,29 +15,50 @@
 #endif
 
 
+// CHECK-LABEL: @test_svset3_bf16_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset3_bf16_014svbfloat16x3_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
 svbfloat16x3_t test_svset3_bf16_0(svbfloat16x3_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset3_bf16_0
-  // CHECK: %[[INSERT:.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> %tuple, i32 0, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 24 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset3_bf16'}}
   return SVE_ACLE_FUNC(svset3,_bf16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset3_bf16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset3_bf16_114svbfloat16x3_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
 svbfloat16x3_t test_svset3_bf16_1(svbfloat16x3_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset3_bf16_1
-  // CHECK: %[[INSERT:.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> %tuple, i32 1, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 24 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset3_bf16'}}
   return SVE_ACLE_FUNC(svset3,_bf16,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset3_bf16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset3_bf16_214svbfloat16x3_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> [[TMP0]]
+//
 svbfloat16x3_t test_svset3_bf16_2(svbfloat16x3_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset3_bf16_2
-  // CHECK: %[[INSERT:.*]] = call <vscale x 24 x bfloat> @llvm.aarch64.sve.tuple.set.nxv24bf16.nxv8bf16(<vscale x 24 x bfloat> %tuple, i32 2, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 24 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset3_bf16'}}
   return SVE_ACLE_FUNC(svset3,_bf16,,)(tuple, 2, x);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
index df9e06a72ba44..d815a44fd4a63 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set3.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -15,90 +16,167 @@
 
 // NOTE: For these tests clang converts the struct parameter into
 // several parameters, one for each member of the original struct.
+// CHECK-LABEL: @test_svset3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svset3_s810svint8x3_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
 svint8x3_t test_svset3_s8(svint8x3_t tuple, svint8_t x)
 {
-  // CHECK-LABEL: test_svset3_s8
-  // CHECK: %[[INSERT:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> %tuple, i32 1, <vscale x 16 x i8> %x)
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_s8,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_s1611svint16x3_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 2, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
 svint16x3_t test_svset3_s16(svint16x3_t tuple, svint16_t x)
 {
-  // CHECK-LABEL: test_svset3_s16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> %tuple, i32 2, <vscale x 8 x i16> %x)
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_s16,,)(tuple, 2, x);
 }
 
+// CHECK-LABEL: @test_svset3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_s3211svint32x3_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 0, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
 svint32x3_t test_svset3_s32(svint32x3_t tuple, svint32_t x)
 {
-  // CHECK-LABEL: test_svset3_s32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> %tuple, i32 0, <vscale x 4 x i32> %x)
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_s32,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_s6411svint64x3_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
 svint64x3_t test_svset3_s64(svint64x3_t tuple, svint64_t x)
 {
-  // CHECK-LABEL: test_svset3_s64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> %tuple, i32 1, <vscale x 2 x i64> %x)
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_s64,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svset3_u811svuint8x3_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> [[TUPLE:%.*]], i32 2, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> [[TMP0]]
+//
 svuint8x3_t test_svset3_u8(svuint8x3_t tuple, svuint8_t x)
 {
-  // CHECK-LABEL: test_svset3_u8
-  // CHECK: %[[INSERT:.*]] = call <vscale x 48 x i8> @llvm.aarch64.sve.tuple.set.nxv48i8.nxv16i8(<vscale x 48 x i8> %tuple, i32 2, <vscale x 16 x i8> %x)
-  // CHECK-NEXT: ret <vscale x 48 x i8> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_u8,,)(tuple, 2, x);
 }
 
+// CHECK-LABEL: @test_svset3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_u1612svuint16x3_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> [[TUPLE:%.*]], i32 0, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> [[TMP0]]
+//
 svuint16x3_t test_svset3_u16(svuint16x3_t tuple, svuint16_t x)
 {
-  // CHECK-LABEL: test_svset3_u16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 24 x i16> @llvm.aarch64.sve.tuple.set.nxv24i16.nxv8i16(<vscale x 24 x i16> %tuple, i32 0, <vscale x 8 x i16> %x)
-  // CHECK-NEXT: ret <vscale x 24 x i16> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_u16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_u3212svuint32x3_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> [[TMP0]]
+//
 svuint32x3_t test_svset3_u32(svuint32x3_t tuple, svuint32_t x)
 {
-  // CHECK-LABEL: test_svset3_u32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 12 x i32> @llvm.aarch64.sve.tuple.set.nxv12i32.nxv4i32(<vscale x 12 x i32> %tuple, i32 1, <vscale x 4 x i32> %x)
-  // CHECK-NEXT: ret <vscale x 12 x i32> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_u32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_u6412svuint64x3_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> [[TUPLE:%.*]], i32 2, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> [[TMP0]]
+//
 svuint64x3_t test_svset3_u64(svuint64x3_t tuple, svuint64_t x)
 {
-  // CHECK-LABEL: test_svset3_u64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 6 x i64> @llvm.aarch64.sve.tuple.set.nxv6i64.nxv2i64(<vscale x 6 x i64> %tuple, i32 2, <vscale x 2 x i64> %x)
-  // CHECK-NEXT: ret <vscale x 6 x i64> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_u64,,)(tuple, 2, x);
 }
 
+// CHECK-LABEL: @test_svset3_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.tuple.set.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 0, <vscale x 8 x half> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_f1613svfloat16x3_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.tuple.set.nxv24f16.nxv8f16(<vscale x 24 x half> [[TUPLE:%.*]], i32 0, <vscale x 8 x half> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 24 x half> [[TMP0]]
+//
 svfloat16x3_t test_svset3_f16(svfloat16x3_t tuple, svfloat16_t x)
 {
-  // CHECK-LABEL: test_svset3_f16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 24 x half> @llvm.aarch64.sve.tuple.set.nxv24f16.nxv8f16(<vscale x 24 x half> %tuple, i32 0, <vscale x 8 x half> %x)
-  // CHECK-NEXT: ret <vscale x 24 x half> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_f16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.tuple.set.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_f3213svfloat32x3_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.tuple.set.nxv12f32.nxv4f32(<vscale x 12 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 12 x float> [[TMP0]]
+//
 svfloat32x3_t test_svset3_f32(svfloat32x3_t tuple, svfloat32_t x)
 {
-  // CHECK-LABEL: test_svset3_f32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 12 x float> @llvm.aarch64.sve.tuple.set.nxv12f32.nxv4f32(<vscale x 12 x float> %tuple, i32 1, <vscale x 4 x float> %x)
-  // CHECK-NEXT: ret <vscale x 12 x float> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_f32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset3_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.tuple.set.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2, <vscale x 2 x double> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset3_f6413svfloat64x3_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.tuple.set.nxv6f64.nxv2f64(<vscale x 6 x double> [[TUPLE:%.*]], i32 2, <vscale x 2 x double> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 6 x double> [[TMP0]]
+//
 svfloat64x3_t test_svset3_f64(svfloat64x3_t tuple, svfloat64_t x)
 {
-  // CHECK-LABEL: test_svset3_f64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 6 x double> @llvm.aarch64.sve.tuple.set.nxv6f64.nxv2f64(<vscale x 6 x double> %tuple, i32 2, <vscale x 2 x double> %x)
-  // CHECK-NEXT: ret <vscale x 6 x double> %[[INSERT]]
   return SVE_ACLE_FUNC(svset3,_f64,,)(tuple, 2, x);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
index 9f8466db1359a..c7a758faa1d78 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -14,38 +15,66 @@
 #endif
 
 
+// CHECK-LABEL: @test_svset4_bf16_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset4_bf16_014svbfloat16x4_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 0, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
 svbfloat16x4_t test_svset4_bf16_0(svbfloat16x4_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset4_bf16_0
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> %tuple, i32 0, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset4_bf16'}}
   return SVE_ACLE_FUNC(svset4,_bf16,,)(tuple, 0, x);
 }
 
+// CHECK-LABEL: @test_svset4_bf16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset4_bf16_114svbfloat16x4_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 1, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
 svbfloat16x4_t test_svset4_bf16_1(svbfloat16x4_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset4_bf16_1
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> %tuple, i32 1, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset4_bf16'}}
   return SVE_ACLE_FUNC(svset4,_bf16,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_bf16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset4_bf16_214svbfloat16x4_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 2, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
 svbfloat16x4_t test_svset4_bf16_2(svbfloat16x4_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset4_bf16_2
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> %tuple, i32 2, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset4_bf16'}}
   return SVE_ACLE_FUNC(svset4,_bf16,,)(tuple, 2, x);
 }
 
+// CHECK-LABEL: @test_svset4_bf16_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3, <vscale x 8 x bfloat> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svset4_bf16_314svbfloat16x4_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> [[TUPLE:%.*]], i32 3, <vscale x 8 x bfloat> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
 svbfloat16x4_t test_svset4_bf16_3(svbfloat16x4_t tuple, svbfloat16_t x)
 {
-  // CHECK-LABEL: test_svset4_bf16_3
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x bfloat> @llvm.aarch64.sve.tuple.set.nxv32bf16.nxv8bf16(<vscale x 32 x bfloat> %tuple, i32 3, <vscale x 8 x bfloat> %x)
-  // CHECK-NEXT: ret <vscale x 32 x bfloat> %[[INSERT]]
   // expected-warning at +1 {{implicit declaration of function 'svset4_bf16'}}
   return SVE_ACLE_FUNC(svset4,_bf16,,)(tuple, 3, x);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
index 7d5cc547991f5..729284b0fc7c4 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_set4.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -13,90 +14,167 @@
 #endif
 
 
+// CHECK-LABEL: @test_svset4_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svset4_s810svint8x4_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 1, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
 svint8x4_t test_svset4_s8(svint8x4_t tuple, svint8_t x)
 {
-  // CHECK-LABEL: test_svset4_s8
-  // CHECK: %[[INSERT:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> %tuple, i32 1, <vscale x 16 x i8> %x)
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_s8,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_s1611svint16x4_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 3, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
 svint16x4_t test_svset4_s16(svint16x4_t tuple, svint16_t x)
 {
-  // CHECK-LABEL: test_svset4_s16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> %tuple, i32 3, <vscale x 8 x i16> %x)
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_s16,,)(tuple, 3, x);
 }
 
+// CHECK-LABEL: @test_svset4_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_s3211svint32x4_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
 svint32x4_t test_svset4_s32(svint32x4_t tuple, svint32_t x)
 {
-  // CHECK-LABEL: test_svset4_s32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> %tuple, i32 1, <vscale x 4 x i32> %x)
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_s32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_s6411svint64x4_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 1, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
 svint64x4_t test_svset4_s64(svint64x4_t tuple, svint64_t x)
 {
-  // CHECK-LABEL: test_svset4_s64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> %tuple, i32 1, <vscale x 2 x i64> %x)
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_s64,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 3, <vscale x 16 x i8> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svset4_u811svuint8x4_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> [[TUPLE:%.*]], i32 3, <vscale x 16 x i8> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
+//
 svuint8x4_t test_svset4_u8(svuint8x4_t tuple, svuint8_t x)
 {
-  // CHECK-LABEL: test_svset4_u8
-  // CHECK: %[[INSERT:.*]] = call <vscale x 64 x i8> @llvm.aarch64.sve.tuple.set.nxv64i8.nxv16i8(<vscale x 64 x i8> %tuple, i32 3, <vscale x 16 x i8> %x)
-  // CHECK-NEXT: ret <vscale x 64 x i8> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_u8,,)(tuple, 3, x);
 }
 
+// CHECK-LABEL: @test_svset4_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 1, <vscale x 8 x i16> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_u1612svuint16x4_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> [[TUPLE:%.*]], i32 1, <vscale x 8 x i16> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
 svuint16x4_t test_svset4_u16(svuint16x4_t tuple, svuint16_t x)
 {
-  // CHECK-LABEL: test_svset4_u16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x i16> @llvm.aarch64.sve.tuple.set.nxv32i16.nxv8i16(<vscale x 32 x i16> %tuple, i32 1, <vscale x 8 x i16> %x)
-  // CHECK-NEXT: ret <vscale x 32 x i16> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_u16,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_u3212svuint32x4_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> [[TUPLE:%.*]], i32 1, <vscale x 4 x i32> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
 svuint32x4_t test_svset4_u32(svuint32x4_t tuple, svuint32_t x)
 {
-  // CHECK-LABEL: test_svset4_u32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x i32> @llvm.aarch64.sve.tuple.set.nxv16i32.nxv4i32(<vscale x 16 x i32> %tuple, i32 1, <vscale x 4 x i32> %x)
-  // CHECK-NEXT: ret <vscale x 16 x i32> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_u32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3, <vscale x 2 x i64> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_u6412svuint64x4_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> [[TUPLE:%.*]], i32 3, <vscale x 2 x i64> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
 svuint64x4_t test_svset4_u64(svuint64x4_t tuple, svuint64_t x)
 {
-  // CHECK-LABEL: test_svset4_u64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 8 x i64> @llvm.aarch64.sve.tuple.set.nxv8i64.nxv2i64(<vscale x 8 x i64> %tuple, i32 3, <vscale x 2 x i64> %x)
-  // CHECK-NEXT: ret <vscale x 8 x i64> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_u64,,)(tuple, 3, x);
 }
 
+// CHECK-LABEL: @test_svset4_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.tuple.set.nxv32f16.nxv8f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_f1613svfloat16x4_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.tuple.set.nxv32f16.nxv8f16(<vscale x 32 x half> [[TUPLE:%.*]], i32 1, <vscale x 8 x half> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 32 x half> [[TMP0]]
+//
 svfloat16x4_t test_svset4_f16(svfloat16x4_t tuple, svfloat16_t x)
 {
-  // CHECK-LABEL: test_svset4_f16
-  // CHECK: %[[INSERT:.*]] = call <vscale x 32 x half> @llvm.aarch64.sve.tuple.set.nxv32f16.nxv8f16(<vscale x 32 x half> %tuple, i32 1, <vscale x 8 x half> %x)
-  // CHECK-NEXT: ret <vscale x 32 x half> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_f16,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.tuple.set.nxv16f32.nxv4f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_f3213svfloat32x4_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.tuple.set.nxv16f32.nxv4f32(<vscale x 16 x float> [[TUPLE:%.*]], i32 1, <vscale x 4 x float> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
 svfloat32x4_t test_svset4_f32(svfloat32x4_t tuple, svfloat32_t x)
 {
-  // CHECK-LABEL: test_svset4_f32
-  // CHECK: %[[INSERT:.*]] = call <vscale x 16 x float> @llvm.aarch64.sve.tuple.set.nxv16f32.nxv4f32(<vscale x 16 x float> %tuple, i32 1, <vscale x 4 x float> %x)
-  // CHECK-NEXT: ret <vscale x 16 x float> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_f32,,)(tuple, 1, x);
 }
 
+// CHECK-LABEL: @test_svset4_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.tuple.set.nxv8f64.nxv2f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 3, <vscale x 2 x double> [[X:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svset4_f6413svfloat64x4_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.tuple.set.nxv8f64.nxv2f64(<vscale x 8 x double> [[TUPLE:%.*]], i32 3, <vscale x 2 x double> [[X:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x double> [[TMP0]]
+//
 svfloat64x4_t test_svset4_f64(svfloat64x4_t tuple, svfloat64_t x)
 {
-  // CHECK-LABEL: test_svset4_f64
-  // CHECK: %[[INSERT:.*]] = call <vscale x 8 x double> @llvm.aarch64.sve.tuple.set.nxv8f64.nxv2f64(<vscale x 8 x double> %tuple, i32 3, <vscale x 2 x double> %x)
-  // CHECK-NEXT: ret <vscale x 8 x double> %[[INSERT]]
   return SVE_ACLE_FUNC(svset4,_f64,,)(tuple, 3, x);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_setffr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_setffr.c
index d97743195d0fa..9e561b4770d43 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_setffr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_setffr.c
@@ -1,13 +1,21 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svsetffr(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.setffr()
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svsetffrv(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.setffr()
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svsetffr()
 {
-  // CHECK-LABEL: test_svsetffr
-  // CHECK: call void @llvm.aarch64.sve.setffr()
-  // CHECK: ret void
   svsetffr();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice-bfloat.c
index f58b90fd255b7..99294b92ec2fe 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice-bfloat.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,12 +14,20 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsplice_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.splice.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsplice_bf16u10__SVBool_tu14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.splice.nxv8bf16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP1]]
+//
 svbfloat16_t test_svsplice_bf16(svbool_t pg, svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsplice_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.splice.nxv8bf16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svsplice_bf16'}}
   return SVE_ACLE_FUNC(svsplice,_bf16,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice.c
index 3e0fb5c129e28..7086eddf56c9f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_splice.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,99 +14,185 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsplice_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.splice.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsplice_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.splice.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsplice_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsplice_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.splice.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.splice.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.splice.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsplice_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsplice_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.splice.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.splice.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_s32u10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.splice.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsplice_s32(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsplice_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.splice.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_s32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.splice.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_s64u10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.splice.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsplice_s64(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsplice_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.splice.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_s64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.splice.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsplice_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.splice.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsplice_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsplice_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.splice.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.splice.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.splice.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsplice_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsplice_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.splice.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_u16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.splice.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.splice.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsplice_u32(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsplice_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.splice.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_u32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.splice.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.splice.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsplice_u64(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsplice_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.splice.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_u64,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.splice.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_f16u10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.splice.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsplice_f16(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsplice_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.splice.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_f16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.splice.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_f32u10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.splice.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsplice_f32(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsplice_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.splice.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_f32,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsplice_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.splice.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsplice_f64u10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.splice.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsplice_f64(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsplice_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.splice.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsplice,_f64,,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sqrt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sqrt.c
index a089b34e344a7..c8d04b65d75f8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sqrt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sqrt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,83 +14,155 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsqrt_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsqrt_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsqrt_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsqrt_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f16_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsqrt_f16_m(svfloat16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f32_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsqrt_f32_m(svfloat32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f64_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsqrt_f64_m(svfloat64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsqrt_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsqrt.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsqrt_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsqrt.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svsqrt_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqrt_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsqrt_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svsqrt_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsqrt.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsqrt,_f64,_x,)(pg, op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1-bfloat.c
index f3a9381e556f0..f41816e2c3b0b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,24 +14,42 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z15test_svst1_bf16u10__SVBool_tPu6__bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_bf16(svbool_t pg, bfloat16_t *base, svbfloat16_t data)
 {
-  // CHECK-LABEL: test_svst1_bf16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8bf16(<vscale x 8 x bfloat> %data, <vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK: ret void
   // expected-warning at +1 {{implicit declaration of function 'svst1_bf16'}}
   return SVE_ACLE_FUNC(svst1,_bf16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z20test_svst1_vnum_bf16u10__SVBool_tPu6__bf16lu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_bf16(svbool_t pg, bfloat16_t *base, int64_t vnum, svbfloat16_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8bf16(<vscale x 8 x bfloat> %data, <vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK: ret void
   // expected-warning at +1 {{implicit declaration of function 'svst1_vnum_bf16'}}
   return SVE_ACLE_FUNC(svst1_vnum,_bf16,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1.c
index 4929d9c97cfcc..af3a711935739 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,601 +14,1138 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst1_s8u10__SVBool_tPau10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_s8(svbool_t pg, int8_t *base, svint8_t data)
 {
-  // CHECK-LABEL: test_svst1_s8
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_s8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_s16u10__SVBool_tPsu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_s16(svbool_t pg, int16_t *base, svint16_t data)
 {
-  // CHECK-LABEL: test_svst1_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_s16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_s32u10__SVBool_tPiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_s32(svbool_t pg, int32_t *base, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_s32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_s64u10__SVBool_tPlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_s64(svbool_t pg, int64_t *base, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_s64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst1_u8u10__SVBool_tPhu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_u8(svbool_t pg, uint8_t *base, svuint8_t data)
 {
-  // CHECK-LABEL: test_svst1_u8
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_u8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_u16u10__SVBool_tPtu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_u16(svbool_t pg, uint16_t *base, svuint16_t data)
 {
-  // CHECK-LABEL: test_svst1_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_u16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_u32u10__SVBool_tPju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_u32(svbool_t pg, uint32_t *base, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_u32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_u64u10__SVBool_tPmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_u64(svbool_t pg, uint64_t *base, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_u64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_f16u10__SVBool_tPDhu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_f16(svbool_t pg, float16_t *base, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svst1_f16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8f16(<vscale x 8 x half> %data, <vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_f16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_f32u10__SVBool_tPfu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_f32(svbool_t pg, float32_t *base, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_f32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst1_f64u10__SVBool_tPdu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_f64(svbool_t pg, float64_t *base, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1,_f64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst1_vnum_s8u10__SVBool_tPalu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_s8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_s8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_s16u10__SVBool_tPslu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_s16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_s32u10__SVBool_tPilu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_s32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_s64u10__SVBool_tPllu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_s64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst1_vnum_u8u10__SVBool_tPhlu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_u8
-  // CHECK: %[[BASE:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_u8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_u16u10__SVBool_tPtlu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_u16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_u32u10__SVBool_tPjlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_u32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_u64u10__SVBool_tPmlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_u64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_f16u10__SVBool_tPDhlu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv8f16(<vscale x 8 x half> %data, <vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_f16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_f32u10__SVBool_tPflu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_vnum,_f32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst1_vnum_f64u10__SVBool_tPdlu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_vnum_f64(svbool_t pg, float64_t *base, int64_t vnum, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_vnum_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[BASE:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BASE]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.st1.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %[[GEP]])
   return SVE_ACLE_FUNC(svst1_vnum,_f64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z30test_svst1_scatter_u32base_s32u10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_s32(svbool_t pg, svuint32_t bases, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,,_s32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z30test_svst1_scatter_u64base_s64u10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_s64(svbool_t pg, svuint64_t bases, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,,_s64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z30test_svst1_scatter_u32base_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_u32(svbool_t pg, svuint32_t bases, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,,_u32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z30test_svst1_scatter_u64base_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_u64(svbool_t pg, svuint64_t bases, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,,_u64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z30test_svst1_scatter_u32base_f32u10__SVBool_tu12__SVUint32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_f32(svbool_t pg, svuint32_t bases, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,,_f32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z30test_svst1_scatter_u64base_f64u10__SVBool_tu12__SVUint64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_f64(svbool_t pg, svuint64_t bases, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,,_f64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_s32offset_s32u10__SVBool_tPiu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s32offset_s32(svbool_t pg, int32_t *base, svint32_t offsets, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s32,offset,_s32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_s64offset_s64u10__SVBool_tPlu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s64offset_s64(svbool_t pg, int64_t *base, svint64_t offsets, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s64,offset,_s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_s32offset_u32u10__SVBool_tPju11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s32offset_u32(svbool_t pg, uint32_t *base, svint32_t offsets, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s32,offset,_u32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_s64offset_u64u10__SVBool_tPmu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s64offset_u64(svbool_t pg, uint64_t *base, svint64_t offsets, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s64,offset,_u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_s32offset_f32u10__SVBool_tPfu11__SVInt32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s32offset_f32(svbool_t pg, float32_t *base, svint32_t offsets, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s32offset_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.sxtw.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s32,offset,_f32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_s64offset_f64u10__SVBool_tPdu11__SVInt64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s64offset_f64(svbool_t pg, float64_t *base, svint64_t offsets, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s64offset_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s64,offset,_f64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_u32offset_s32u10__SVBool_tPiu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32offset_s32(svbool_t pg, int32_t *base, svuint32_t offsets, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u32,offset,_s32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_u64offset_s64u10__SVBool_tPlu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64offset_s64(svbool_t pg, int64_t *base, svuint64_t offsets, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u64,offset,_s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_u32offset_u32u10__SVBool_tPju12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32offset_u32(svbool_t pg, uint32_t *base, svuint32_t offsets, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u32,offset,_u32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_u64offset_u64u10__SVBool_tPmu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64offset_u64(svbool_t pg, uint64_t *base, svuint64_t offsets, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u64,offset,_u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_u32offset_f32u10__SVBool_tPfu12__SVUint32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32offset_f32(svbool_t pg, float32_t *base, svuint32_t offsets, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32offset_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.uxtw.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u32,offset,_f32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svst1_scatter_u64offset_f64u10__SVBool_tPdu12__SVUint64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64offset_f64(svbool_t pg, float64_t *base, svuint64_t offsets, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64offset_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u64,offset,_f64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z37test_svst1_scatter_u32base_offset_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_offset_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,_offset,_s32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z37test_svst1_scatter_u64base_offset_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_offset_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,_offset,_s64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z37test_svst1_scatter_u32base_offset_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_offset_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,_offset,_u32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z37test_svst1_scatter_u64base_offset_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_offset_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,_offset,_u64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z37test_svst1_scatter_u32base_offset_f32u10__SVBool_tu12__SVUint32_tlu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_offset_f32(svbool_t pg, svuint32_t bases, int64_t offset, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_offset_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,_offset,_f32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z37test_svst1_scatter_u64base_offset_f64u10__SVBool_tu12__SVUint64_tlu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_offset_f64(svbool_t pg, svuint64_t bases, int64_t offset, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_offset_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,_offset,_f64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_s32index_s32u10__SVBool_tPiu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s32index_s32(svbool_t pg, int32_t *base, svint32_t indices, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s32,index,_s32)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_s64index_s64u10__SVBool_tPlu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s64index_s64(svbool_t pg, int64_t *base, svint64_t indices, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s64,index,_s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_s32index_u32u10__SVBool_tPju11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s32index_u32(svbool_t pg, uint32_t *base, svint32_t indices, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s32,index,_u32)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_s64index_u64u10__SVBool_tPmu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s64index_u64(svbool_t pg, uint64_t *base, svint64_t indices, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s64,index,_u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s32index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_s32index_f32u10__SVBool_tPfu11__SVInt32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s32index_f32(svbool_t pg, float32_t *base, svint32_t indices, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s32index_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.sxtw.index.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s32,index,_f32)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_s64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_s64index_f64u10__SVBool_tPdu11__SVInt64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_s64index_f64(svbool_t pg, float64_t *base, svint64_t indices, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_s64index_f64
-  // CHECK: %[[PG:.*]]  = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.index.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,s64,index,_f64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_u32index_s32u10__SVBool_tPiu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32index_s32(svbool_t pg, int32_t *base, svuint32_t indices, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32index_s32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u32,index,_s32)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_u64index_s64u10__SVBool_tPlu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64index_s64(svbool_t pg, int64_t *base, svuint64_t indices, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64index_s64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %0, i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u64,index,_s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_u32index_u32u10__SVBool_tPju12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32index_u32(svbool_t pg, uint32_t *base, svuint32_t indices, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32index_u32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u32,index,_u32)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_u64index_u64u10__SVBool_tPmu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64index_u64(svbool_t pg, uint64_t *base, svuint64_t indices, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64index_u64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u64,index,_u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_u32index_f32u10__SVBool_tPfu12__SVUint32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32index_f32(svbool_t pg, float32_t *base, svuint32_t indices, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32index_f32
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.uxtw.index.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %base, <vscale x 4 x i32> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u32,index,_f32)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z31test_svst1_scatter_u64index_f64u10__SVBool_tPdu12__SVUint64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64index_f64(svbool_t pg, float64_t *base, svuint64_t indices, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64index_f64
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.index.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter_,u64,index,_f64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z36test_svst1_scatter_u32base_index_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index, svint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_index_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,_index,_s32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z36test_svst1_scatter_u64base_index_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index, svint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_index_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,_index,_s64)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z36test_svst1_scatter_u32base_index_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index, svuint32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_index_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,_index,_u32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z36test_svst1_scatter_u64base_index_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index, svuint64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_index_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,_index,_u64)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u32base_index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z36test_svst1_scatter_u32base_index_f32u10__SVBool_tu12__SVUint32_tlu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u32base_index_f32(svbool_t pg, svuint32_t bases, int64_t index, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u32base_index_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %bases, i64 %[[SHL]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u32base,_index,_f32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svst1_scatter_u64base_index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z36test_svst1_scatter_u64base_index_f64u10__SVBool_tu12__SVUint64_tlu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst1_scatter_u64base_index_f64(svbool_t pg, svuint64_t bases, int64_t index, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svst1_scatter_u64base_index_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SHL:.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.st1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %bases, i64 %[[SHL]])
-  // CHECK: ret void
   return SVE_ACLE_FUNC(svst1_scatter,_u64base,_index,_f64)(pg, bases, index, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1b.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1b.c
index dc10c5fbf5a1f..c14f4eeeb58e9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1b.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1b.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s 2>&1 | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s 2>&1 | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s |& FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s |& FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1h.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1h.c
index edf36e373160d..70094af2960d8 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1h.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1h.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s 2>&1 | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s 2>&1 | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s |& FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s |& FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1w.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1w.c
index cce9b35771126..f688303047f5f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1w.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st1w.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s 2>&1 | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s 2>&1 | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s |& FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - -emit-llvm %s |& FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
index a1480b4c7d0f5..f9c334dfb2734 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -11,26 +12,48 @@
 #else
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
+// CHECK-LABEL: @test_svst2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z15test_svst2_bf16u10__SVBool_tPu6__bf1614svbfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_bf16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_bf16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z20test_svst2_vnum_bf16u10__SVBool_tPu6__bf16l14svbfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_bf16(svbool_t pg, bfloat16_t *base, int64_t vnum, svbfloat16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_bf16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_bf16,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
index 30e882652f06c..58c8e6d189b3c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st2.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,262 +14,500 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst2_s8u10__SVBool_tPa10svint8x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_s8(svbool_t pg, int8_t *base, svint8x2_t data)
 {
-  // CHECK-LABEL: test_svst2_s8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 1)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_s8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_s16u10__SVBool_tPs11svint16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_s16(svbool_t pg, int16_t *base, svint16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_s16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_s16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_s32u10__SVBool_tPi11svint32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_s32(svbool_t pg, int32_t *base, svint32x2_t data)
 {
-  // CHECK-LABEL: test_svst2_s32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_s32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_s64u10__SVBool_tPl11svint64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_s64(svbool_t pg, int64_t *base, svint64x2_t data)
 {
-  // CHECK-LABEL: test_svst2_s64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_s64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst2_u8u10__SVBool_tPh11svuint8x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_u8(svbool_t pg, uint8_t *base, svuint8x2_t data)
 {
-  // CHECK-LABEL: test_svst2_u8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 1)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_u8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_u16u10__SVBool_tPt12svuint16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_u16(svbool_t pg, uint16_t *base, svuint16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_u16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_u16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_u32u10__SVBool_tPj12svuint32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_u32(svbool_t pg, uint32_t *base, svuint32x2_t data)
 {
-  // CHECK-LABEL: test_svst2_u32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_u32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_u64u10__SVBool_tPm12svuint64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_u64(svbool_t pg, uint64_t *base, svuint64x2_t data)
 {
-  // CHECK-LABEL: test_svst2_u64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_u64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_f16u10__SVBool_tPDh13svfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_f16(svbool_t pg, float16_t *base, svfloat16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_f16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_f16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_f32u10__SVBool_tPf13svfloat32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_f32(svbool_t pg, float32_t *base, svfloat32x2_t data)
 {
-  // CHECK-LABEL: test_svst2_f32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_f32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst2_f64u10__SVBool_tPd13svfloat64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_f64(svbool_t pg, float64_t *base, svfloat64x2_t data)
 {
-  // CHECK-LABEL: test_svst2_f64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2,_f64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst2_vnum_s8u10__SVBool_tPal10svint8x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_s8
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 1)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_s8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_s16u10__SVBool_tPsl11svint16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_s16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_s16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_s32u10__SVBool_tPil11svint32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_s32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_s32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_s64u10__SVBool_tPll11svint64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_s64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_s64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst2_vnum_u8u10__SVBool_tPhl11svuint8x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_u8
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 1)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_u8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_u16u10__SVBool_tPtl12svuint16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_u16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_u16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_u32u10__SVBool_tPjl12svuint32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_u32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_u32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_u64u10__SVBool_tPml12svuint64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_u64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_u64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_f16u10__SVBool_tPDhl13svfloat16x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_f16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_f16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_f32u10__SVBool_tPfl13svfloat32x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_f32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_f32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst2_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst2_vnum_f64u10__SVBool_tPdl13svfloat64x2_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst2_vnum_f64(svbool_t pg, float64_t *base, int64_t vnum, svfloat64x2_t data)
 {
-  // CHECK-LABEL: test_svst2_vnum_f64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %data, i32 1)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st2.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst2_vnum,_f64,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
index 5a9e11a8dedf4..3e6e96b706699 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,28 +13,52 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst3_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z15test_svst3_bf16u10__SVBool_tPu6__bf1614svbfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_bf16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x bfloat> %[[V2]], <vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_bf16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z20test_svst3_vnum_bf16u10__SVBool_tPu6__bf16l14svbfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_bf16(svbool_t pg, bfloat16_t *base, int64_t vnum, svbfloat16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_bf16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv24bf16(<vscale x 24 x bfloat> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x bfloat> %[[V2]], <vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_bf16,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
index b2ad6e23fef09..e55d9bc653924 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st3.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,284 +14,544 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst3_s8u10__SVBool_tPa10svint8x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_s8(svbool_t pg, int8_t *base, svint8x3_t data)
 {
-  // CHECK-LABEL: test_svst3_s8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 2)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_s8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_s16u10__SVBool_tPs11svint16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_s16(svbool_t pg, int16_t *base, svint16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_s16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_s16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_s32u10__SVBool_tPi11svint32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_s32(svbool_t pg, int32_t *base, svint32x3_t data)
 {
-  // CHECK-LABEL: test_svst3_s32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_s32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_s64u10__SVBool_tPl11svint64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_s64(svbool_t pg, int64_t *base, svint64x3_t data)
 {
-  // CHECK-LABEL: test_svst3_s64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_s64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst3_u8u10__SVBool_tPh11svuint8x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_u8(svbool_t pg, uint8_t *base, svuint8x3_t data)
 {
-  // CHECK-LABEL: test_svst3_u8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 2)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_u8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_u16u10__SVBool_tPt12svuint16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_u16(svbool_t pg, uint16_t *base, svuint16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_u16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_u16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_u32u10__SVBool_tPj12svuint32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_u32(svbool_t pg, uint32_t *base, svuint32x3_t data)
 {
-  // CHECK-LABEL: test_svst3_u32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_u32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_u64u10__SVBool_tPm12svuint64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_u64(svbool_t pg, uint64_t *base, svuint64x3_t data)
 {
-  // CHECK-LABEL: test_svst3_u64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_u64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_f16u10__SVBool_tPDh13svfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_f16(svbool_t pg, float16_t *base, svfloat16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_f16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x half> %[[V2]], <vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_f16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_f32u10__SVBool_tPf13svfloat32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_f32(svbool_t pg, float32_t *base, svfloat32x3_t data)
 {
-  // CHECK-LABEL: test_svst3_f32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x float> %[[V2]], <vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_f32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst3_f64u10__SVBool_tPd13svfloat64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_f64(svbool_t pg, float64_t *base, svfloat64x3_t data)
 {
-  // CHECK-LABEL: test_svst3_f64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x double> %[[V2]], <vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3,_f64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst3_vnum_s8u10__SVBool_tPal10svint8x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_s8
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 2)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_s8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_s16u10__SVBool_tPsl11svint16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_s16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_s16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_s32u10__SVBool_tPil11svint32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_s32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_s32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_s64u10__SVBool_tPll11svint64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_s64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_s64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst3_vnum_u8u10__SVBool_tPhl11svuint8x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_u8
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv48i8(<vscale x 48 x i8> %data, i32 2)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_u8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_u16u10__SVBool_tPtl12svuint16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_u16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv24i16(<vscale x 24 x i16> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_u16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_u32u10__SVBool_tPjl12svuint32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_u32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv12i32(<vscale x 12 x i32> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_u32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_u64u10__SVBool_tPml12svuint64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_u64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv6i64(<vscale x 6 x i64> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_u64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_f16u10__SVBool_tPDhl13svfloat16x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_f16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv24f16(<vscale x 24 x half> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x half> %[[V2]], <vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_f16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_f32u10__SVBool_tPfl13svfloat32x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_f32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv12f32(<vscale x 12 x float> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x float> %[[V2]], <vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_f32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst3_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst3_vnum_f64u10__SVBool_tPdl13svfloat64x3_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst3_vnum_f64(svbool_t pg, float64_t *base, int64_t vnum, svfloat64x3_t data)
 {
-  // CHECK-LABEL: test_svst3_vnum_f64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv6f64(<vscale x 6 x double> %data, i32 2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st3.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x double> %[[V2]], <vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst3_vnum,_f64,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
index 3c8bb7da46386..46b93b14564b3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,30 +13,56 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst4_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z15test_svst4_bf16u10__SVBool_tPu6__bf1614svbfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x bfloat> [[TMP2]], <vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_bf16(svbool_t pg, bfloat16_t *base, svbfloat16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 3)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x bfloat> %[[V2]], <vscale x 8 x bfloat> %[[V3]], <vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_bf16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x bfloat> [[TMP6]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z20test_svst4_vnum_bf16u10__SVBool_tPu6__bf16l14svbfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> [[TMP3]], <vscale x 8 x bfloat> [[TMP4]], <vscale x 8 x bfloat> [[TMP5]], <vscale x 8 x bfloat> [[TMP6]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_bf16(svbool_t pg, bfloat16_t *base, int64_t vnum, svbfloat16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_bf16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv32bf16(<vscale x 32 x bfloat> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x bfloat> %[[V2]], <vscale x 8 x bfloat> %[[V3]], <vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_bf16,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
index 69b12103ee15d..1374c0952dabe 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_st4.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,306 +14,588 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svst4_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst4_s8u10__SVBool_tPa10svint8x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_s8(svbool_t pg, int8_t *base, svint8x4_t data)
 {
-  // CHECK-LABEL: test_svst4_s8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 3)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i8> %[[V3]], <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_s8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_s16u10__SVBool_tPs11svint16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_s16(svbool_t pg, int16_t *base, svint16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_s16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i16> %[[V3]], <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_s16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_s32u10__SVBool_tPi11svint32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_s32(svbool_t pg, int32_t *base, svint32x4_t data)
 {
-  // CHECK-LABEL: test_svst4_s32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i32> %[[V3]], <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_s32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_s64u10__SVBool_tPl11svint64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_s64(svbool_t pg, int64_t *base, svint64x4_t data)
 {
-  // CHECK-LABEL: test_svst4_s64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i64> %[[V3]], <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_s64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z13test_svst4_u8u10__SVBool_tPh11svuint8x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_u8(svbool_t pg, uint8_t *base, svuint8x4_t data)
 {
-  // CHECK-LABEL: test_svst4_u8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 3)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i8> %[[V3]], <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_u8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_u16u10__SVBool_tPt12svuint16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_u16(svbool_t pg, uint16_t *base, svuint16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_u16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i16> %[[V3]], <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_u16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_u32u10__SVBool_tPj12svuint32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_u32(svbool_t pg, uint32_t *base, svuint32x4_t data)
 {
-  // CHECK-LABEL: test_svst4_u32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i32> %[[V3]], <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_u32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_u64u10__SVBool_tPm12svuint64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_u64(svbool_t pg, uint64_t *base, svuint64x4_t data)
 {
-  // CHECK-LABEL: test_svst4_u64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i64> %[[V3]], <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_u64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_f16u10__SVBool_tPDh13svfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_f16(svbool_t pg, float16_t *base, svfloat16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 3)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x half> %[[V2]], <vscale x 8 x half> %[[V3]], <vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_f16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_f32u10__SVBool_tPf13svfloat32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_f32(svbool_t pg, float32_t *base, svfloat32x4_t data)
 {
-  // CHECK-LABEL: test_svst4_f32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x float> %[[V2]], <vscale x 4 x float> %[[V3]], <vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_f32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z14test_svst4_f64u10__SVBool_tPd13svfloat64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_f64(svbool_t pg, float64_t *base, svfloat64x4_t data)
 {
-  // CHECK-LABEL: test_svst4_f64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x double> %[[V2]], <vscale x 2 x double> %[[V3]], <vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4,_f64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst4_vnum_s8u10__SVBool_tPal10svint8x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_s8
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 3)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i8> %[[V3]], <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_s8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_s16u10__SVBool_tPsl11svint16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_s16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i16> %[[V3]], <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_s16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_s32u10__SVBool_tPil11svint32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_s32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i32> %[[V3]], <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_s32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_s64u10__SVBool_tPll11svint64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_s64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i64> %[[V3]], <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_s64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z18test_svst4_vnum_u8u10__SVBool_tPhl11svuint8x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> [[TMP2]], <vscale x 16 x i8> [[TMP3]], <vscale x 16 x i8> [[TMP4]], <vscale x 16 x i8> [[TMP5]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_u8
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv64i8(<vscale x 64 x i8> %data, i32 3)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %[[V2]], <vscale x 16 x i8> %[[V3]], <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_u8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_u16u10__SVBool_tPtl12svuint16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> [[TMP3]], <vscale x 8 x i16> [[TMP4]], <vscale x 8 x i16> [[TMP5]], <vscale x 8 x i16> [[TMP6]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_u16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv32i16(<vscale x 32 x i16> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %[[V2]], <vscale x 8 x i16> %[[V3]], <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_u16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_u32u10__SVBool_tPjl12svuint32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> [[TMP3]], <vscale x 4 x i32> [[TMP4]], <vscale x 4 x i32> [[TMP5]], <vscale x 4 x i32> [[TMP6]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_u32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv16i32(<vscale x 16 x i32> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %[[V2]], <vscale x 4 x i32> %[[V3]], <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_u32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_u64u10__SVBool_tPml12svuint64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> [[TMP3]], <vscale x 2 x i64> [[TMP4]], <vscale x 2 x i64> [[TMP5]], <vscale x 2 x i64> [[TMP6]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_u64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv8i64(<vscale x 8 x i64> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %[[V2]], <vscale x 2 x i64> %[[V3]], <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_u64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x half> [[TMP6]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_f16u10__SVBool_tPDhl13svfloat16x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> [[TMP3]], <vscale x 8 x half> [[TMP4]], <vscale x 8 x half> [[TMP5]], <vscale x 8 x half> [[TMP6]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_f16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv32f16(<vscale x 32 x half> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x half> %[[V2]], <vscale x 8 x half> %[[V3]], <vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_f16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x float> [[TMP6]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_f32u10__SVBool_tPfl13svfloat32x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> [[TMP3]], <vscale x 4 x float> [[TMP4]], <vscale x 4 x float> [[TMP5]], <vscale x 4 x float> [[TMP6]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_f32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv16f32(<vscale x 16 x float> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x float> %[[V2]], <vscale x 4 x float> %[[V3]], <vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_f32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svst4_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
+// CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x double> [[TMP6]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z19test_svst4_vnum_f64u10__SVBool_tPdl13svfloat64x4_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP4:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 2)
+// CPP-CHECK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> [[DATA]], i32 3)
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> [[TMP3]], <vscale x 2 x double> [[TMP4]], <vscale x 2 x double> [[TMP5]], <vscale x 2 x double> [[TMP6]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svst4_vnum_f64(svbool_t pg, float64_t *base, int64_t vnum, svfloat64x4_t data)
 {
-  // CHECK-LABEL: test_svst4_vnum_f64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 1)
-  // CHECK-DAG: %[[V2:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 2)
-  // CHECK-DAG: %[[V3:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv8f64(<vscale x 8 x double> %data, i32 3)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.st4.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x double> %[[V2]], <vscale x 2 x double> %[[V3]], <vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svst4_vnum,_f64,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1-bfloat.c
index 05fb09613489b..7964219cce010 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,24 +14,42 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svstnt1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z17test_svstnt1_bf16u10__SVBool_tPu6__bf16u14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_bf16(svbool_t pg, bfloat16_t *base, svbfloat16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_bf16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8bf16(<vscale x 8 x bfloat> %data, <vscale x 8 x i1> %[[PG]], bfloat* %base)
-  // CHECK-NEXT: ret
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_bf16'}}
   return SVE_ACLE_FUNC(svstnt1,_bf16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z22test_svstnt1_vnum_bf16u10__SVBool_tPu6__bf16lu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast bfloat* [[BASE:%.*]] to <vscale x 8 x bfloat>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], bfloat* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_bf16(svbool_t pg, bfloat16_t *base, int64_t vnum, svbfloat16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_bf16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8bf16(<vscale x 8 x bfloat> %data, <vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
-  // CHECK-NEXT: ret
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_vnum_bf16'}}
   return SVE_ACLE_FUNC(svstnt1_vnum,_bf16,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1.c
index 9dc6770fa7aca..6ccc7f9f696db 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_stnt1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 
 #include <arm_sve.h>
 
@@ -14,218 +15,412 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svstnt1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z15test_svstnt1_s8u10__SVBool_tPau10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_s8(svbool_t pg, int8_t *base, svint8_t data)
 {
-  // CHECK-LABEL: test_svstnt1_s8
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_s8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_s16u10__SVBool_tPsu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_s16(svbool_t pg, int16_t *base, svint16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_s16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_s16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_s32u10__SVBool_tPiu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_s32(svbool_t pg, int32_t *base, svint32_t data)
 {
-  // CHECK-LABEL: test_svstnt1_s32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_s32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_s64u10__SVBool_tPlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_s64(svbool_t pg, int64_t *base, svint64_t data)
 {
-  // CHECK-LABEL: test_svstnt1_s64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_s64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z15test_svstnt1_u8u10__SVBool_tPhu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_u8(svbool_t pg, uint8_t *base, svuint8_t data)
 {
-  // CHECK-LABEL: test_svstnt1_u8
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_u8,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_u16u10__SVBool_tPtu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_u16(svbool_t pg, uint16_t *base, svuint16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_u16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_u16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_u32u10__SVBool_tPju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_u32(svbool_t pg, uint32_t *base, svuint32_t data)
 {
-  // CHECK-LABEL: test_svstnt1_u32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_u32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_u64u10__SVBool_tPmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_u64(svbool_t pg, uint64_t *base, svuint64_t data)
 {
-  // CHECK-LABEL: test_svstnt1_u64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_u64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_f16u10__SVBool_tPDhu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_f16(svbool_t pg, float16_t *base, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_f16
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8f16(<vscale x 8 x half> %data, <vscale x 8 x i1> %[[PG]], half* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_f16,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_f32u10__SVBool_tPfu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_f32(svbool_t pg, float32_t *base, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svstnt1_f32
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_f32,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z16test_svstnt1_f64u10__SVBool_tPdu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_f64(svbool_t pg, float64_t *base, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svstnt1_f64
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %base)
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1,_f64,,)(pg, base, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z20test_svstnt1_vnum_s8u10__SVBool_tPalu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_s8(svbool_t pg, int8_t *base, int64_t vnum, svint8_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_s8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_s8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_s16u10__SVBool_tPslu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_s16(svbool_t pg, int16_t *base, int64_t vnum, svint16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_s16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_s16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_s32u10__SVBool_tPilu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_s32(svbool_t pg, int32_t *base, int64_t vnum, svint32_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_s32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_s32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_s64u10__SVBool_tPllu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_s64(svbool_t pg, int64_t *base, int64_t vnum, svint64_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_s64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_s64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z20test_svstnt1_vnum_u8u10__SVBool_tPhlu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[BASE:%.*]] to <vscale x 16 x i8>*
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* [[TMP0]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i1> [[PG:%.*]], i8* [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_u8(svbool_t pg, uint8_t *base, int64_t vnum, svuint8_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_u8
-  // CHECK: %[[BITCAST:.*]] = bitcast i8* %base to <vscale x 16 x i8>*
-  // CHECK: %[[GEP:.*]] = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i1> %pg, i8* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_u8,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_u16u10__SVBool_tPtlu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i16* [[BASE:%.*]] to <vscale x 8 x i16>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], i16* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_u16(svbool_t pg, uint16_t *base, int64_t vnum, svuint16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_u16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i16* %base to <vscale x 8 x i16>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i1> %[[PG]], i16* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_u16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_u32u10__SVBool_tPjlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i32* [[BASE:%.*]] to <vscale x 4 x i32>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_u32(svbool_t pg, uint32_t *base, int64_t vnum, svuint32_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_u32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i32* %base to <vscale x 4 x i32>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> %[[PG]], i32* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_u32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_u64u10__SVBool_tPmlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast i64* [[BASE:%.*]] to <vscale x 2 x i64>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_u64(svbool_t pg, uint64_t *base, int64_t vnum, svuint64_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_u64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast i64* %base to <vscale x 2 x i64>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> %[[PG]], i64* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_u64,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_f16u10__SVBool_tPDhlu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast half* [[BASE:%.*]] to <vscale x 8 x half>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i1> [[TMP0]], half* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_f16(svbool_t pg, float16_t *base, int64_t vnum, svfloat16_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_f16
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast half* %base to <vscale x 8 x half>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv8f16(<vscale x 8 x half> %data, <vscale x 8 x i1> %[[PG]], half* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_f16,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_f32u10__SVBool_tPflu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[BASE:%.*]] to <vscale x 4 x float>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_f32(svbool_t pg, float32_t *base, int64_t vnum, svfloat32_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_f32
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast float* %base to <vscale x 4 x float>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> %[[PG]], float* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_f32,,)(pg, base, vnum, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_vnum_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z21test_svstnt1_vnum_f64u10__SVBool_tPdlu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast double* [[BASE:%.*]] to <vscale x 2 x double>*
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* [[TMP1]], i64 [[VNUM:%.*]], i64 0
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_vnum_f64(svbool_t pg, float64_t *base, int64_t vnum, svfloat64_t data)
 {
-  // CHECK-LABEL: test_svstnt1_vnum_f64
-  // CHECK-DAG: %[[BITCAST:.*]] = bitcast double* %base to <vscale x 2 x double>*
-  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %[[BITCAST]], i64 %vnum, i64 0
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> %[[PG]], double* %[[GEP]])
-  // CHECK-NEXT: ret
   return SVE_ACLE_FUNC(svstnt1_vnum,_f64,,)(pg, base, vnum, data);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sub.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sub.c
index 977c7937d0ebf..a6e1c7016ac61 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sub.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sub.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,639 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsub_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsub_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsub_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svsub_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svsub_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svsub_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsub_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsub_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsub_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsub_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsub_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsub_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsub_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsub_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsub_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsub_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsub_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsub_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsub_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsub_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsub_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsub_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsub_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsub_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsub_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsub_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsub_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsub_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsub_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsub_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsub_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsub_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svsub_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svsub_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svsub_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svsub_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsub_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svsub_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svsub_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svsub_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svsub_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsub_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsub_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svsub_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svsub_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svsub_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsub_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsub_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsub_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsub_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsub_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsub_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsub_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svsub_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svsub_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svsub_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsub_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsub_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsub_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsub_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsub_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svsub_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsub_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svsub_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsub_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svsub_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsub_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsub_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsub_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsub_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsub_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsub_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsub_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsub_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsub_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsub_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsub_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsub_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsub_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsub_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svsub_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svsub_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svsub_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svsub_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svsub_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svsub_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svsub_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svsub_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsub_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsub_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svsub_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svsub_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsub,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_subr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_subr.c
index 95dadd7cae1b5..218d43dcedf3d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_subr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_subr.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,639 +14,1210 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsubr_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubr_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsubr_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svsubr_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svsubr_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svsubr_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubr_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsubr_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsubr_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsubr_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsubr_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubr_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsubr_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubr_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubr_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsubr_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubr_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsubr_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsubr_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsubr_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsubr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubr_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsubr_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubr_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubr_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsubr_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubr_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsubr_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsubr_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsubr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsubr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubr_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svsubr_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svsubr_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svsubr_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svsubr_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubr_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svsubr_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svsubr_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svsubr_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svsubr_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubr_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsubr_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svsubr_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svsubr_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svsubr_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubr_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsubr_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsubr_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsubr_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsubr_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubr_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsubr_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svsubr_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svsubr_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svsubr_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubr_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsubr_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsubr_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsubr_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsubr_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.subr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f16_zu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svsubr_f16_z(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f32_zu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svsubr_f32_z(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f64_zu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svsubr_f64_z(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsubr_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsubr_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsubr_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svsubr_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svsubr_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubr_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svsubr_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_f64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f16_zu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[TMP2]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP3]]
+//
 svfloat16_t test_svsubr_n_f16_z(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.sel.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %[[SEL]], <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f32_zu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[TMP2]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP3]]
+//
 svfloat32_t test_svsubr_n_f32_z(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.sel.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %[[SEL]], <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f64_zu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[TMP2]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP3]]
+//
 svfloat64_t test_svsubr_n_f64_z(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.sel.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %[[SEL]], <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f16_mu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svsubr_n_f16_m(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f32_mu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svsubr_n_f32_m(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f64_mu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svsubr_n_f64_m(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f16_xu10__SVBool_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svsubr_n_f16_x(svbool_t pg, svfloat16_t op1, float16_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fsubr.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f32_xu10__SVBool_tu13__SVFloat32_tf(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svsubr_n_f32_x(svbool_t pg, svfloat32_t op1, float32_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.dup.x.nxv4f32(float %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fsubr.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubr_n_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubr_n_f64_xu10__SVBool_tu13__SVFloat64_td(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svsubr_n_f64_x(svbool_t pg, svfloat64_t op1, float64_t op2)
 {
-  // CHECK-LABEL: test_svsubr_n_f64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.dup.x.nxv2f64(double %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fsubr.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %[[DUP]])
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svsubr,_n_f64,_x,)(pg, op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sudot.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sudot.c
index 6f0e4daaa2d6c..e23bf6ce10d4d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sudot.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_sudot.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,45 +13,88 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsudot_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Z:%.*]], <vscale x 16 x i8> [[Y:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsudot_s32u11__SVInt32_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Z:%.*]], <vscale x 16 x i8> [[Y:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsudot_s32(svint32_t x, svint8_t y, svuint8_t z) {
-  // CHECK-LABEL: test_svsudot_s32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %z, <vscale x 16 x i8> %y)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svsudot, _s32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_svsudot_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[Z:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[Y:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsudot_n_s32u11__SVInt32_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[Z:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[Y:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsudot_n_s32(svint32_t x, svint8_t y, uint8_t z) {
-  // CHECK-LABEL: test_svsudot_n_s32
-  // CHECK: %[[SPLAT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %z)
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %[[SPLAT]], <vscale x 16 x i8> %y)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svsudot, _n_s32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_svsudot_lane_s32_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svsudot_lane_s32_0u11__SVInt32_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsudot_lane_s32_0(svint32_t x, svint8_t y, svuint8_t z) {
-  // CHECK-LABEL: test_svsudot_lane_s32_0
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svsudot_lane, _s32, , )(x, y, z, 0);
 }
 
+// CHECK-LABEL: @test_svsudot_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svsudot_lane_s32_1u11__SVInt32_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsudot_lane_s32_1(svint32_t x, svint8_t y, svuint8_t z) {
-  // CHECK-LABEL: test_svsudot_lane_s32_1
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svsudot_lane, _s32, , )(x, y, z, 1);
 }
 
+// CHECK-LABEL: @test_svsudot_lane_s32_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svsudot_lane_s32_2u11__SVInt32_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsudot_lane_s32_2(svint32_t x, svint8_t y, svuint8_t z) {
-  // CHECK-LABEL: test_svsudot_lane_s32_2
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 2)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svsudot_lane, _s32, , )(x, y, z, 2);
 }
 
+// CHECK-LABEL: @test_svsudot_lane_s32_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svsudot_lane_s32_3u11__SVInt32_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsudot_lane_s32_3(svint32_t x, svint8_t y, svuint8_t z) {
-  // CHECK-LABEL: test_svsudot_lane_s32_3
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sudot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svsudot_lane, _s32, , )(x, y, z, 3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl-bfloat.c
index a4a4424925631..934749a651b3d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl-bfloat.c
@@ -1,10 +1,11 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -14,10 +15,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtbl_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl_bf16u14__SVBFloat16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl.nxv8bf16(<vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svtbl_bf16(svbfloat16_t data, svuint16_t indices) {
-  // CHECK-LABEL: test_svtbl_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl.nxv8bf16(<vscale x 8 x bfloat> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svtbl_bf16'}}
   return SVE_ACLE_FUNC(svtbl, _bf16, , )(data, indices);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl.c
index 66be572885fe3..ac61b64117e83 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tbl.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,90 +14,167 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtbl_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svtbl_s8u10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svtbl_s8(svint8_t data, svuint8_t indices)
 {
-  // CHECK-LABEL: test_svtbl_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i8> %indices)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_s8,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_s16u11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svtbl_s16(svint16_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbl_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_s16,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_s32u11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svtbl_s32(svint32_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbl_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_s32,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_s64u11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svtbl_s64(svint64_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbl_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_s64,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svtbl_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svtbl_u8(svuint8_t data, svuint8_t indices)
 {
-  // CHECK-LABEL: test_svtbl_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl.nxv16i8(<vscale x 16 x i8> %data, <vscale x 16 x i8> %indices)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_u8,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svtbl_u16(svuint16_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbl_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl.nxv8i16(<vscale x 8 x i16> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_u16,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svtbl_u32(svuint32_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbl_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_u32,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svtbl_u64(svuint64_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbl_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_u64,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_f16u13__SVFloat16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl.nxv8f16(<vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtbl_f16(svfloat16_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbl_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl.nxv8f16(<vscale x 8 x half> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_f16,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_f32u13__SVFloat32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtbl_f32(svfloat32_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbl_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_f32,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl_f64u13__SVFloat64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtbl_f64(svfloat64_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbl_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtbl,_f64,,)(data, indices);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tmad.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tmad.c
index f06deba60c7ab..9e5d99a5ff547 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tmad.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tmad.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,34 +14,62 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtmad_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftmad.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtmad_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftmad.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtmad_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svtmad_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftmad.x.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtmad,_f16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svtmad_f16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftmad.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svtmad_f16_1u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftmad.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtmad_f16_1(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svtmad_f16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftmad.x.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2, i32 7)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtmad,_f16,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svtmad_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftmad.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtmad_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftmad.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtmad_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svtmad_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftmad.x.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtmad,_f32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svtmad_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftmad.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtmad_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftmad.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtmad_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svtmad_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftmad.x.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtmad,_f64,,)(op1, op2, 0);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-bfloat.c
index 36862a268049f..4ac4d0ee8d4a4 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn1.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtrn1_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn1.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svtrn1_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn1.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svtrn1_bf16'}}
   return SVE_ACLE_FUNC(svtrn1,_bf16,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64-bfloat.c
index 3814ae8673048..c2d77cf7ca134 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn1q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtrn1_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn1q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svtrn1_bf16(svbfloat16_t op1, svbfloat16_t op2) {
-  // CHECK-LABEL: test_svtrn1_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn1q.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svtrn1q_bf16'}}
   return SVE_ACLE_FUNC(svtrn1q, _bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64.c
index d4753ced350ef..2ae77daac9e8c 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1-fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,79 +13,156 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn1_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svtrn1_s8(svint8_t op1, svint8_t op2) {
-  // CHECK-LABEL: test_svtrn1_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _s8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svtrn1_s16(svint16_t op1, svint16_t op2) {
-  // CHECK-LABEL: test_svtrn1_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _s16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svtrn1_s32(svint32_t op1, svint32_t op2) {
-  // CHECK-LABEL: test_svtrn1_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _s32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svtrn1_s64(svint64_t op1, svint64_t op2) {
-  // CHECK-LABEL: test_svtrn1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _s64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn1_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svtrn1_u8(svuint8_t op1, svuint8_t op2) {
-  // CHECK-LABEL: test_svtrn1_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _u8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svtrn1_u16(svuint16_t op1, svuint16_t op2) {
-  // CHECK-LABEL: test_svtrn1_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _u16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svtrn1_u32(svuint32_t op1, svuint32_t op2) {
-  // CHECK-LABEL: test_svtrn1_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _u32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svtrn1_u64(svuint64_t op1, svuint64_t op2) {
-  // CHECK-LABEL: test_svtrn1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _u64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn1q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn1q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtrn1_f16(svfloat16_t op1, svfloat16_t op2) {
-  // CHECK-LABEL: test_svtrn1_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn1q.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _f16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn1q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn1q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtrn1_f32(svfloat32_t op1, svfloat32_t op2) {
-  // CHECK-LABEL: test_svtrn1_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn1q.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _f32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn1q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn1q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtrn1_f64(svfloat64_t op1, svfloat64_t op2) {
-  // CHECK-LABEL: test_svtrn1_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn1q.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1q, _f64, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1.c
index f847984e3e157..e96ff2fd258ef 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,131 +14,245 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn1_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svtrn1_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svtrn1_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svtrn1_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svtrn1_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn1_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svtrn1_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn1.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svtrn1_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn1.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svtrn1_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn1.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svtrn1_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn1.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn1.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn1.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtrn1_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn1.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn1.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn1.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtrn1_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn1.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn1.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn1.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtrn1_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn1.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn1,_f64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.trn1.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn1_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.trn1.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svtrn1_b8(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.trn1.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svtrn1_b8(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.trn1.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.trn1.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svtrn1_b16(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_b16
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.trn1.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svtrn1_b16(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.trn1.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.trn1.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svtrn1_b32(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_b32
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.trn1.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svtrn1_b32(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn1_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.trn1.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn1_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.trn1.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svtrn1_b64(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn1_b64
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.trn1.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svtrn1_b64(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-bfloat.c
index a23b5af5f506a..ce77ba79fecb6 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn2.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtrn2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn2.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svtrn2_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn2.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svtrn2_bf16'}}
   return SVE_ACLE_FUNC(svtrn2,_bf16,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64-bfloat.c
index 0ce57ce58aaac..3d12cb3586070 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn2q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtrn2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn2q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svtrn2_bf16(svbfloat16_t op1, svbfloat16_t op2) {
-  // CHECK-LABEL: test_svtrn2_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.trn2q.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svtrn2q_bf16'}}
   return SVE_ACLE_FUNC(svtrn2q, _bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64.c
index bdec8cbf0af06..c119178410f43 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2-fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,79 +13,156 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svtrn2_s8(svint8_t op1, svint8_t op2) {
-  // CHECK-LABEL: test_svtrn2_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _s8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svtrn2_s16(svint16_t op1, svint16_t op2) {
-  // CHECK-LABEL: test_svtrn2_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _s16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svtrn2_s32(svint32_t op1, svint32_t op2) {
-  // CHECK-LABEL: test_svtrn2_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _s32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svtrn2_s64(svint64_t op1, svint64_t op2) {
-  // CHECK-LABEL: test_svtrn2_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _s64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svtrn2_u8(svuint8_t op1, svuint8_t op2) {
-  // CHECK-LABEL: test_svtrn2_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _u8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svtrn2_u16(svuint16_t op1, svuint16_t op2) {
-  // CHECK-LABEL: test_svtrn2_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _u16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svtrn2_u32(svuint32_t op1, svuint32_t op2) {
-  // CHECK-LABEL: test_svtrn2_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _u32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svtrn2_u64(svuint64_t op1, svuint64_t op2) {
-  // CHECK-LABEL: test_svtrn2_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _u64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn2q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn2q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtrn2_f16(svfloat16_t op1, svfloat16_t op2) {
-  // CHECK-LABEL: test_svtrn2_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn2q.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _f16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn2q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn2q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtrn2_f32(svfloat32_t op1, svfloat32_t op2) {
-  // CHECK-LABEL: test_svtrn2_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn2q.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _f32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn2q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn2q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtrn2_f64(svfloat64_t op1, svfloat64_t op2) {
-  // CHECK-LABEL: test_svtrn2_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn2q.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2q, _f64, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2.c
index f45e4f5daf894..c4b3dab2a1291 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_trn2.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,131 +14,245 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtrn2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svtrn2_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svtrn2_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svtrn2_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svtrn2_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svtrn2_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.trn2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svtrn2_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.trn2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svtrn2_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.trn2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svtrn2_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.trn2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtrn2_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.trn2.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtrn2_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.trn2.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtrn2_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.trn2.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtrn2,_f64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtrn2_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svtrn2_b8(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.trn2.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svtrn2_b8(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.trn2.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.trn2.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svtrn2_b16(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_b16
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.trn2.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svtrn2_b16(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.trn2.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.trn2.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svtrn2_b32(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_b32
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.trn2.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svtrn2_b32(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtrn2_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.trn2.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtrn2_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.trn2.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svtrn2_b64(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svtrn2_b64
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.trn2.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svtrn2_b64(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tsmul.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tsmul.c
index 80b59b8d1e42f..3971b9043b7d0 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tsmul.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tsmul.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtsmul_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftsmul.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtsmul_f16u13__SVFloat16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftsmul.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtsmul_f16(svfloat16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svtsmul_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftsmul.x.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtsmul,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtsmul_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftsmul.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtsmul_f32u13__SVFloat32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftsmul.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtsmul_f32(svfloat32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svtsmul_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftsmul.x.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtsmul,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtsmul_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftsmul.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtsmul_f64u13__SVFloat64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftsmul.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtsmul_f64(svfloat64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svtsmul_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftsmul.x.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtsmul,_f64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tssel.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tssel.c
index 33a789f41bb8b..07fce87d646f3 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tssel.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_tssel.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,26 +14,47 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtssel_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftssel.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtssel_f16u13__SVFloat16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftssel.x.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtssel_f16(svfloat16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svtssel_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ftssel.x.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtssel,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtssel_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftssel.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtssel_f32u13__SVFloat32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftssel.x.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtssel_f32(svfloat32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svtssel_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ftssel.x.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtssel,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svtssel_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftssel.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtssel_f64u13__SVFloat64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftssel.x.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtssel_f64(svfloat64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svtssel_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ftssel.x.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svtssel,_f64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef-bfloat.c
index 47c7fb7bbcf27..6bacab64b374d 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef-bfloat.c
@@ -1,13 +1,20 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef_bf16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> undef
+//
 svbfloat16_t test_svundef_bf16()
 {
-  // CHECK-LABEL: test_svundef_bf16
-  // CHECK: ret <vscale x 8 x bfloat> undef
   // expected-warning at +1 {{implicit declaration of function 'svundef_bf16'}}
   return svundef_bf16();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef.c
index 5095c47557cbf..70804395e3bbc 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef.c
@@ -1,82 +1,149 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z15test_svundef_s8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> undef
+//
 svint8_t test_svundef_s8()
 {
-  // CHECK-LABEL: test_svundef_s8
-  // CHECK: ret <vscale x 16 x i8> undef
   return svundef_s8();
 }
 
+// CHECK-LABEL: @test_svundef_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_s16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> undef
+//
 svint16_t test_svundef_s16()
 {
-  // CHECK-LABEL: test_svundef_s16
-  // CHECK: ret <vscale x 8 x i16> undef
   return svundef_s16();
 }
 
+// CHECK-LABEL: @test_svundef_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_s32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> undef
+//
 svint32_t test_svundef_s32()
 {
-  // CHECK-LABEL: test_svundef_s32
-  // CHECK: ret <vscale x 4 x i32> undef
   return svundef_s32();
 }
 
+// CHECK-LABEL: @test_svundef_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_s64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> undef
+//
 svint64_t test_svundef_s64()
 {
-  // CHECK-LABEL: test_svundef_s64
-  // CHECK: ret <vscale x 2 x i64> undef
   return svundef_s64();
 }
 
+// CHECK-LABEL: @test_svundef_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z15test_svundef_u8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> undef
+//
 svuint8_t test_svundef_u8()
 {
-  // CHECK-LABEL: test_svundef_u8
-  // CHECK: ret <vscale x 16 x i8> undef
   return svundef_u8();
 }
 
+// CHECK-LABEL: @test_svundef_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_u16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> undef
+//
 svuint16_t test_svundef_u16()
 {
-  // CHECK-LABEL: test_svundef_u16
-  // CHECK: ret <vscale x 8 x i16> undef
   return svundef_u16();
 }
 
+// CHECK-LABEL: @test_svundef_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_u32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> undef
+//
 svuint32_t test_svundef_u32()
 {
-  // CHECK-LABEL: test_svundef_u32
-  // CHECK: ret <vscale x 4 x i32> undef
   return svundef_u32();
 }
 
+// CHECK-LABEL: @test_svundef_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_u64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> undef
+//
 svuint64_t test_svundef_u64()
 {
-  // CHECK-LABEL: test_svundef_u64
-  // CHECK: ret <vscale x 2 x i64> undef
   return svundef_u64();
 }
 
+// CHECK-LABEL: @test_svundef_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x half> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_f16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> undef
+//
 svfloat16_t test_svundef_f16()
 {
-  // CHECK-LABEL: test_svundef_f16
-  // CHECK: ret <vscale x 8 x half> undef
   return svundef_f16();
 }
 
+// CHECK-LABEL: @test_svundef_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x float> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_f32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> undef
+//
 svfloat32_t test_svundef_f32()
 {
-  // CHECK-LABEL: test_svundef_f32
-  // CHECK: ret <vscale x 4 x float> undef
   return svundef_f32();
 }
 
+// CHECK-LABEL: @test_svundef_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 2 x double> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef_f64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> undef
+//
 svfloat64_t test_svundef_f64()
 {
-  // CHECK-LABEL: test_svundef_f64
-  // CHECK: ret <vscale x 2 x double> undef
   return svundef_f64();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2-bfloat.c
index 62fb4f8f75aab..923c24cb08ce6 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2-bfloat.c
@@ -1,13 +1,20 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x bfloat> undef
+//
+// CPP-CHECK-LABEL: @_Z18test_svundef2_bf16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x bfloat> undef
+//
 svbfloat16x2_t test_svundef2_bf16()
 {
-  // CHECK-LABEL: test_svundef2_bf16
-  // CHECK: ret <vscale x 16 x bfloat> undef
   // expected-warning at +1 {{implicit declaration of function 'svundef2_bf16'}}
   return svundef2_bf16();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2.c
index e928b3c218f9a..652295cba15bf 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef2.c
@@ -1,81 +1,148 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 32 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef2_s8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> undef
+//
 svint8x2_t test_svundef2_s8()
 {
-  // CHECK-LABEL: test_svundef2_s8
-  // CHECK: ret <vscale x 32 x i8> undef
   return svundef2_s8();
 }
 
+// CHECK-LABEL: @test_svundef2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_s16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> undef
+//
 svint16x2_t test_svundef2_s16()
 {
-  // CHECK-LABEL: test_svundef2_s16
-  // CHECK: ret <vscale x 16 x i16> undef
   return svundef2_s16();
 }
 
+// CHECK-LABEL: @test_svundef2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_s32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> undef
+//
 svint32x2_t test_svundef2_s32()
 {
-  // CHECK-LABEL: test_svundef2_s32
-  // CHECK: ret <vscale x 8 x i32> undef
   return svundef2_s32();
 }
 
+// CHECK-LABEL: @test_svundef2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_s64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> undef
+//
 svint64x2_t test_svundef2_s64()
 {
-  // CHECK-LABEL: test_svundef2_s64
-  // CHECK: ret <vscale x 4 x i64> undef
   return svundef2_s64();
 }
 
+// CHECK-LABEL: @test_svundef2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 32 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef2_u8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i8> undef
+//
 svuint8x2_t test_svundef2_u8()
 {
-  // CHECK-LABEL: test_svundef2_u8
-  // CHECK: ret <vscale x 32 x i8> undef
   return svundef2_u8();
 }
 
+// CHECK-LABEL: @test_svundef2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_u16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i16> undef
+//
 svuint16x2_t test_svundef2_u16()
 {
-  // CHECK-LABEL: test_svundef2_u16
-  // CHECK: ret <vscale x 16 x i16> undef
   return svundef2_u16();
 }
 
+// CHECK-LABEL: @test_svundef2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_u32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i32> undef
+//
 svuint32x2_t test_svundef2_u32()
 {
-  // CHECK-LABEL: test_svundef2_u32
-  // CHECK: ret <vscale x 8 x i32> undef
   return svundef2_u32();
 }
 
+// CHECK-LABEL: @test_svundef2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_u64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i64> undef
+//
 svuint64x2_t test_svundef2_u64()
 {
-  // CHECK-LABEL: test_svundef2_u64
-  // CHECK: ret <vscale x 4 x i64> undef
   return svundef2_u64();
 }
 
+// CHECK-LABEL: @test_svundef2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x half> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_f16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x half> undef
+//
 svfloat16x2_t test_svundef2_f16()
 {
-  // CHECK-LABEL: test_svundef2_f16
-  // CHECK: ret <vscale x 16 x half> undef
   return svundef2_f16();
 }
 
+// CHECK-LABEL: @test_svundef2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x float> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_f32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x float> undef
+//
 svfloat32x2_t test_svundef2_f32()
 {
-  // CHECK-LABEL: test_svundef2_f32
-  // CHECK: ret <vscale x 8 x float> undef
   return svundef2_f32();
 }
 
+// CHECK-LABEL: @test_svundef2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 4 x double> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef2_f64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 4 x double> undef
+//
 svfloat64x2_t test_svundef2_f64()
 {
-  // CHECK-LABEL: test_svundef2_f64
-  // CHECK: ret <vscale x 4 x double> undef
   return svundef2_f64();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3-bfloat.c
index 434c180b59097..f5124fdea2580 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3-bfloat.c
@@ -1,13 +1,20 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef3_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 24 x bfloat> undef
+//
+// CPP-CHECK-LABEL: @_Z18test_svundef3_bf16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 24 x bfloat> undef
+//
 svbfloat16x3_t test_svundef3_bf16()
 {
-  // CHECK-LABEL: test_svundef3_bf16
-  // CHECK: ret <vscale x 24 x bfloat> undef
   // expected-warning at +1 {{implicit declaration of function 'svundef3_bf16'}}
   return svundef3_bf16();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3.c
index 00165062585d2..d9c3126be1ced 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef3.c
@@ -1,81 +1,148 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 48 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef3_s8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> undef
+//
 svint8x3_t test_svundef3_s8()
 {
-  // CHECK-LABEL: test_svundef3_s8
-  // CHECK: ret <vscale x 48 x i8> undef
   return svundef3_s8();
 }
 
+// CHECK-LABEL: @test_svundef3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 24 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_s16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> undef
+//
 svint16x3_t test_svundef3_s16()
 {
-  // CHECK-LABEL: test_svundef3_s16
-  // CHECK: ret <vscale x 24 x i16> undef
   return svundef3_s16();
 }
 
+// CHECK-LABEL: @test_svundef3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 12 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_s32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> undef
+//
 svint32x3_t test_svundef3_s32()
 {
-  // CHECK-LABEL: test_svundef3_s32
-  // CHECK: ret <vscale x 12 x i32> undef
   return svundef3_s32();
 }
 
+// CHECK-LABEL: @test_svundef3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 6 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_s64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> undef
+//
 svint64x3_t test_svundef3_s64()
 {
-  // CHECK-LABEL: test_svundef3_s64
-  // CHECK: ret <vscale x 6 x i64> undef
   return svundef3_s64();
 }
 
+// CHECK-LABEL: @test_svundef3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 48 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef3_u8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 48 x i8> undef
+//
 svuint8x3_t test_svundef3_u8()
 {
-  // CHECK-LABEL: test_svundef3_u8
-  // CHECK: ret <vscale x 48 x i8> undef
   return svundef3_u8();
 }
 
+// CHECK-LABEL: @test_svundef3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 24 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_u16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 24 x i16> undef
+//
 svuint16x3_t test_svundef3_u16()
 {
-  // CHECK-LABEL: test_svundef3_u16
-  // CHECK: ret <vscale x 24 x i16> undef
   return svundef3_u16();
 }
 
+// CHECK-LABEL: @test_svundef3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 12 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_u32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 12 x i32> undef
+//
 svuint32x3_t test_svundef3_u32()
 {
-  // CHECK-LABEL: test_svundef3_u32
-  // CHECK: ret <vscale x 12 x i32> undef
   return svundef3_u32();
 }
 
+// CHECK-LABEL: @test_svundef3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 6 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_u64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 6 x i64> undef
+//
 svuint64x3_t test_svundef3_u64()
 {
-  // CHECK-LABEL: test_svundef3_u64
-  // CHECK: ret <vscale x 6 x i64> undef
   return svundef3_u64();
 }
 
+// CHECK-LABEL: @test_svundef3_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 24 x half> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_f16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 24 x half> undef
+//
 svfloat16x3_t test_svundef3_f16()
 {
-  // CHECK-LABEL: test_svundef3_f16
-  // CHECK: ret <vscale x 24 x half> undef
   return svundef3_f16();
 }
 
+// CHECK-LABEL: @test_svundef3_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 12 x float> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_f32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 12 x float> undef
+//
 svfloat32x3_t test_svundef3_f32()
 {
-  // CHECK-LABEL: test_svundef3_f32
-  // CHECK: ret <vscale x 12 x float> undef
   return svundef3_f32();
 }
 
+// CHECK-LABEL: @test_svundef3_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 6 x double> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef3_f64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 6 x double> undef
+//
 svfloat64x3_t test_svundef3_f64()
 {
-  // CHECK-LABEL: test_svundef3_f64
-  // CHECK: ret <vscale x 6 x double> undef
   return svundef3_f64();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4-bfloat.c
index 07ec64a99d587..91f0264b1d821 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4-bfloat.c
@@ -1,13 +1,20 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef4_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 32 x bfloat> undef
+//
+// CPP-CHECK-LABEL: @_Z18test_svundef4_bf16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 32 x bfloat> undef
+//
 svbfloat16x4_t test_svundef4_bf16()
 {
-  // CHECK-LABEL: test_svundef4_bf16
-  // CHECK: ret <vscale x 32 x bfloat> undef
   // expected-warning at +1 {{implicit declaration of function 'svundef4_bf16'}}
   return svundef4_bf16();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4.c
index 98dea16c8a7db..5e9db39995172 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_undef4.c
@@ -1,81 +1,148 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O2 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svundef4_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 64 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef4_s8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> undef
+//
 svint8x4_t test_svundef4_s8()
 {
-  // CHECK-LABEL: test_svundef4_s8
-  // CHECK: ret <vscale x 64 x i8> undef
   return svundef4_s8();
 }
 
+// CHECK-LABEL: @test_svundef4_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 32 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_s16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> undef
+//
 svint16x4_t test_svundef4_s16()
 {
-  // CHECK-LABEL: test_svundef4_s16
-  // CHECK: ret <vscale x 32 x i16> undef
   return svundef4_s16();
 }
 
+// CHECK-LABEL: @test_svundef4_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_s32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> undef
+//
 svint32x4_t test_svundef4_s32()
 {
-  // CHECK-LABEL: test_svundef4_s32
-  // CHECK: ret <vscale x 16 x i32> undef
   return svundef4_s32();
 }
 
+// CHECK-LABEL: @test_svundef4_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_s64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> undef
+//
 svint64x4_t test_svundef4_s64()
 {
-  // CHECK-LABEL: test_svundef4_s64
-  // CHECK: ret <vscale x 8 x i64> undef
   return svundef4_s64();
 }
 
+// CHECK-LABEL: @test_svundef4_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 64 x i8> undef
+//
+// CPP-CHECK-LABEL: @_Z16test_svundef4_u8v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 64 x i8> undef
+//
 svuint8x4_t test_svundef4_u8()
 {
-  // CHECK-LABEL: test_svundef4_u8
-  // CHECK: ret <vscale x 64 x i8> undef
   return svundef4_u8();
 }
 
+// CHECK-LABEL: @test_svundef4_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 32 x i16> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_u16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 32 x i16> undef
+//
 svuint16x4_t test_svundef4_u16()
 {
-  // CHECK-LABEL: test_svundef4_u16
-  // CHECK: ret <vscale x 32 x i16> undef
   return svundef4_u16();
 }
 
+// CHECK-LABEL: @test_svundef4_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x i32> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_u32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i32> undef
+//
 svuint32x4_t test_svundef4_u32()
 {
-  // CHECK-LABEL: test_svundef4_u32
-  // CHECK: ret <vscale x 16 x i32> undef
   return svundef4_u32();
 }
 
+// CHECK-LABEL: @test_svundef4_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x i64> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_u64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i64> undef
+//
 svuint64x4_t test_svundef4_u64()
 {
-  // CHECK-LABEL: test_svundef4_u64
-  // CHECK: ret <vscale x 8 x i64> undef
   return svundef4_u64();
 }
 
+// CHECK-LABEL: @test_svundef4_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 32 x half> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_f16v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 32 x half> undef
+//
 svfloat16x4_t test_svundef4_f16()
 {
-  // CHECK-LABEL: test_svundef4_f16
-  // CHECK: ret <vscale x 32 x half> undef
   return svundef4_f16();
 }
 
+// CHECK-LABEL: @test_svundef4_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 16 x float> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_f32v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 16 x float> undef
+//
 svfloat32x4_t test_svundef4_f32()
 {
-  // CHECK-LABEL: test_svundef4_f32
-  // CHECK: ret <vscale x 16 x float> undef
   return svundef4_f32();
 }
 
+// CHECK-LABEL: @test_svundef4_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    ret <vscale x 8 x double> undef
+//
+// CPP-CHECK-LABEL: @_Z17test_svundef4_f64v(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    ret <vscale x 8 x double> undef
+//
 svfloat64x4_t test_svundef4_f64()
 {
-  // CHECK-LABEL: test_svundef4_f64
-  // CHECK: ret <vscale x 8 x double> undef
   return svundef4_f64();
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpkhi.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpkhi.c
index c27f26a9ecdb5..39768c06b3017 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpkhi.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpkhi.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,59 +14,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svunpkhi_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sunpkhi.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpkhi_s16u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sunpkhi.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svunpkhi_s16(svint8_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sunpkhi.nxv8i16(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpkhi,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpkhi_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sunpkhi.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpkhi_s32u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sunpkhi.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svunpkhi_s32(svint16_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sunpkhi.nxv4i32(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpkhi,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpkhi_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sunpkhi.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpkhi_s64u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sunpkhi.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svunpkhi_s64(svint32_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sunpkhi.nxv2i64(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpkhi,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpkhi_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uunpkhi.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpkhi_u16u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uunpkhi.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svunpkhi_u16(svuint8_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uunpkhi.nxv8i16(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpkhi,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpkhi_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uunpkhi.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpkhi_u32u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uunpkhi.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svunpkhi_u32(svuint16_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uunpkhi.nxv4i32(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpkhi,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpkhi_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uunpkhi.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpkhi_u64u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uunpkhi.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svunpkhi_u64(svuint32_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uunpkhi.nxv2i64(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpkhi,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpkhi_b(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.punpkhi.nxv16i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svunpkhi_bu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.punpkhi.nxv16i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svunpkhi_b(svbool_t op)
 {
-  // CHECK-LABEL: test_svunpkhi_b
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.punpkhi.nxv16i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svunpkhi,_b,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpklo.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpklo.c
index adac7d9b3818b..23dba2eee4b5b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpklo.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_unpklo.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,59 +14,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svunpklo_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sunpklo.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpklo_s16u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sunpklo.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svunpklo_s16(svint8_t op)
 {
-  // CHECK-LABEL: test_svunpklo_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sunpklo.nxv8i16(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpklo,_s16,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpklo_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sunpklo.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpklo_s32u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sunpklo.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svunpklo_s32(svint16_t op)
 {
-  // CHECK-LABEL: test_svunpklo_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sunpklo.nxv4i32(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpklo,_s32,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpklo_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sunpklo.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpklo_s64u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sunpklo.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svunpklo_s64(svint32_t op)
 {
-  // CHECK-LABEL: test_svunpklo_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sunpklo.nxv2i64(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpklo,_s64,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpklo_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uunpklo.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpklo_u16u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uunpklo.nxv8i16(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svunpklo_u16(svuint8_t op)
 {
-  // CHECK-LABEL: test_svunpklo_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uunpklo.nxv8i16(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpklo,_u16,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpklo_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uunpklo.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpklo_u32u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uunpklo.nxv4i32(<vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svunpklo_u32(svuint16_t op)
 {
-  // CHECK-LABEL: test_svunpklo_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uunpklo.nxv4i32(<vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpklo,_u32,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpklo_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uunpklo.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svunpklo_u64u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uunpklo.nxv2i64(<vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svunpklo_u64(svuint32_t op)
 {
-  // CHECK-LABEL: test_svunpklo_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uunpklo.nxv2i64(<vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svunpklo,_u64,,)(op);
 }
 
+// CHECK-LABEL: @test_svunpklo_b(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.punpklo.nxv16i1(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svunpklo_bu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.punpklo.nxv16i1(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svunpklo_b(svbool_t op)
 {
-  // CHECK-LABEL: test_svunpklo_b
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.punpklo.nxv16i1(<vscale x 16 x i1> %op)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svunpklo,_b,,)(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_usdot.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_usdot.c
index e77f509b22c3f..0eefd34b490b2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_usdot.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_usdot.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +i8mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +i8mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,45 +13,88 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svusdot_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svusdot_s32u11__SVInt32_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svusdot_s32(svint32_t x, svuint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svusdot_s32
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusdot, _s32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_svusdot_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[Z:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svusdot_n_s32u11__SVInt32_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[Z:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svusdot_n_s32(svint32_t x, svuint8_t y, int8_t z) {
-  // CHECK-LABEL: test_svusdot_n_s32
-  // CHECK: %[[SPLAT:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %z)
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %[[SPLAT]])
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusdot, _n_s32, , )(x, y, z);
 }
 
+// CHECK-LABEL: @test_svusdot_lane_s32_0(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svusdot_lane_s32_0u11__SVInt32_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svusdot_lane_s32_0(svint32_t x, svuint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svusdot_lane_s32_0
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusdot_lane, _s32, , )(x, y, z, 0);
 }
 
+// CHECK-LABEL: @test_svusdot_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svusdot_lane_s32_1u11__SVInt32_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svusdot_lane_s32_1(svint32_t x, svuint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svusdot_lane_s32_1
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusdot_lane, _s32, , )(x, y, z, 1);
 }
 
+// CHECK-LABEL: @test_svusdot_lane_s32_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 2)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svusdot_lane_s32_2u11__SVInt32_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svusdot_lane_s32_2(svint32_t x, svuint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svusdot_lane_s32_2
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 2)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusdot_lane, _s32, , )(x, y, z, 2);
 }
 
+// CHECK-LABEL: @test_svusdot_lane_s32_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svusdot_lane_s32_3u11__SVInt32_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> [[X:%.*]], <vscale x 16 x i8> [[Y:%.*]], <vscale x 16 x i8> [[Z:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svusdot_lane_s32_3(svint32_t x, svuint8_t y, svint8_t z) {
-  // CHECK-LABEL: test_svusdot_lane_s32_3
-  // CHECK: %[[RET:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usdot.lane.nxv4i32(<vscale x 4 x i32> %x, <vscale x 16 x i8> %y, <vscale x 16 x i8> %z, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[RET]]
   return SVE_ACLE_FUNC(svusdot_lane, _s32, , )(x, y, z, 3);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-bfloat.c
index 509f8b7efa615..87942966b099b 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp1.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svuzp1_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp1.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svuzp1_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp1.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svuzp1_bf16'}}
   return SVE_ACLE_FUNC(svuzp1,_bf16,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64-bfloat.c
index 838edb1cb0e29..ed78a746216c4 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp1q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svuzp1_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp1q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svuzp1_bf16(svbfloat16_t op1, svbfloat16_t op2) {
-  // CHECK-LABEL: test_svuzp1_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp1q.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svuzp1q_bf16'}}
   return SVE_ACLE_FUNC(svuzp1q, _bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64.c
index b4f1289441f45..ef3a0de8759bf 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1-fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,79 +13,156 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp1_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svuzp1_s8(svint8_t op1, svint8_t op2) {
-  // CHECK-LABEL: test_svuzp1_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _s8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svuzp1_s16(svint16_t op1, svint16_t op2) {
-  // CHECK-LABEL: test_svuzp1_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _s16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svuzp1_s32(svint32_t op1, svint32_t op2) {
-  // CHECK-LABEL: test_svuzp1_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _s32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svuzp1_s64(svint64_t op1, svint64_t op2) {
-  // CHECK-LABEL: test_svuzp1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _s64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp1_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svuzp1_u8(svuint8_t op1, svuint8_t op2) {
-  // CHECK-LABEL: test_svuzp1_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _u8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svuzp1_u16(svuint16_t op1, svuint16_t op2) {
-  // CHECK-LABEL: test_svuzp1_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _u16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svuzp1_u32(svuint32_t op1, svuint32_t op2) {
-  // CHECK-LABEL: test_svuzp1_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _u32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svuzp1_u64(svuint64_t op1, svuint64_t op2) {
-  // CHECK-LABEL: test_svuzp1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _u64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp1q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp1q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svuzp1_f16(svfloat16_t op1, svfloat16_t op2) {
-  // CHECK-LABEL: test_svuzp1_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp1q.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _f16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp1q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp1q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svuzp1_f32(svfloat32_t op1, svfloat32_t op2) {
-  // CHECK-LABEL: test_svuzp1_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp1q.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _f32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp1q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp1q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svuzp1_f64(svfloat64_t op1, svfloat64_t op2) {
-  // CHECK-LABEL: test_svuzp1_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp1q.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1q, _f64, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1.c
index 53a2f3f8567c1..bf3f9a55cad6f 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,131 +14,245 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp1_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svuzp1_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svuzp1_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svuzp1_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svuzp1_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp1_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svuzp1_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp1.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svuzp1_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp1.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svuzp1_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp1.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svuzp1_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp1.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp1.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp1.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svuzp1_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp1.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp1.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp1.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svuzp1_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp1.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp1.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp1.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svuzp1_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp1.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp1,_f64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.uzp1.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp1_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.uzp1.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svuzp1_b8(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.uzp1.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svuzp1_b8(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svuzp1_b16(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_b16
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svuzp1_b16(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.uzp1.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.uzp1.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svuzp1_b32(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_b32
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.uzp1.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svuzp1_b32(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp1_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.uzp1.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp1_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.uzp1.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svuzp1_b64(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp1_b64
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.uzp1.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svuzp1_b64(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-bfloat.c
index 6adb017443c0a..9e324b20beba6 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp2.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svuzp2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp2.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svuzp2_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp2.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svuzp2_bf16'}}
   return SVE_ACLE_FUNC(svuzp2,_bf16,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64-bfloat.c
index 372bc8882319e..b32409b58a924 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp2q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svuzp2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp2q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svuzp2_bf16(svbfloat16_t op1, svbfloat16_t op2) {
-  // CHECK-LABEL: test_svuzp2_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.uzp2q.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svuzp2q_bf16'}}
   return SVE_ACLE_FUNC(svuzp2q, _bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64.c
index fc53015551d15..9b6fc7619e3c9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2-fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,79 +13,156 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svuzp2_s8(svint8_t op1, svint8_t op2) {
-  // CHECK-LABEL: test_svuzp2_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _s8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svuzp2_s16(svint16_t op1, svint16_t op2) {
-  // CHECK-LABEL: test_svuzp2_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _s16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svuzp2_s32(svint32_t op1, svint32_t op2) {
-  // CHECK-LABEL: test_svuzp2_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _s32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svuzp2_s64(svint64_t op1, svint64_t op2) {
-  // CHECK-LABEL: test_svuzp2_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _s64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svuzp2_u8(svuint8_t op1, svuint8_t op2) {
-  // CHECK-LABEL: test_svuzp2_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _u8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svuzp2_u16(svuint16_t op1, svuint16_t op2) {
-  // CHECK-LABEL: test_svuzp2_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _u16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svuzp2_u32(svuint32_t op1, svuint32_t op2) {
-  // CHECK-LABEL: test_svuzp2_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _u32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svuzp2_u64(svuint64_t op1, svuint64_t op2) {
-  // CHECK-LABEL: test_svuzp2_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _u64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp2q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp2q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svuzp2_f16(svfloat16_t op1, svfloat16_t op2) {
-  // CHECK-LABEL: test_svuzp2_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp2q.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _f16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp2q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp2q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svuzp2_f32(svfloat32_t op1, svfloat32_t op2) {
-  // CHECK-LABEL: test_svuzp2_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp2q.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _f32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp2q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp2q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svuzp2_f64(svfloat64_t op1, svfloat64_t op2) {
-  // CHECK-LABEL: test_svuzp2_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp2q.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2q, _f64, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2.c
index d0a56563fc427..e56eedf11dcd5 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_uzp2.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,131 +14,245 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuzp2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svuzp2_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svuzp2_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svuzp2_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svuzp2_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svuzp2_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uzp2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svuzp2_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uzp2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svuzp2_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uzp2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svuzp2_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uzp2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svuzp2_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.uzp2.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svuzp2_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.uzp2.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svuzp2_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.uzp2.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svuzp2,_f64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.uzp2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svuzp2_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.uzp2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svuzp2_b8(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.uzp2.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svuzp2_b8(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.uzp2.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.uzp2.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svuzp2_b16(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_b16
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.uzp2.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svuzp2_b16(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.uzp2.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.uzp2.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svuzp2_b32(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_b32
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.uzp2.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svuzp2_b32(op1, op2);
 }
 
+// CHECK-LABEL: @test_svuzp2_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.uzp2.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svuzp2_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.uzp2.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svuzp2_b64(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svuzp2_b64
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.uzp2.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svuzp2_b64(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilele.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilele.c
index 55c4e526dd05d..da4781a5532d9 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilele.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilele.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,142 +14,266 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilele_b8_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilele.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilele_b8_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilele.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilele_b8_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b8_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilele.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilele_b8,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilele.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b16_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilele.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b16_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b16_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilele.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b16,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilele.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b32_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilele.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b32_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b32_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilele.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b32,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilele.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b64_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilele.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b64_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b64_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilele.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b64,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b8_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilels.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilele_b8_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilels.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilele_b8_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b8_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilels.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilele_b8,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilels.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b16_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilels.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b16_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b16_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilels.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b16,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilels.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b32_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilels.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b32_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b32_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilels.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b32,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilels.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b64_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilels.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b64_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b64_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilels.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b64,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b8_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilele.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilele_b8_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilele.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilele_b8_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b8_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilele.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilele_b8,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilele.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b16_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilele.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b16_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b16_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilele.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b16,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilele.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b32_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilele.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b32_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b32_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilele.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b32,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilele.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b64_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilele.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b64_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b64_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilele.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b64,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b8_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilels.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilele_b8_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilels.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilele_b8_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b8_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilels.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilele_b8,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilels.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b16_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilels.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b16_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b16_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilels.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b16,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilels.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b32_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilels.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b32_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b32_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilels.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b32,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilele_b64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilels.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilele_b64_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilels.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilele_b64_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilele_b64_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilels.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilele_b64,_u64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilelt.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilelt.c
index a8275ad7fdaea..f0adc0c9a1fb2 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilelt.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_whilelt.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,142 +14,266 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilelt_b8_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelt.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilelt_b8_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelt.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilelt_b8_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b8_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelt.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilelt_b8,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelt.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b16_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelt.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b16_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b16_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelt.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b16,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelt.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b32_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelt.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b32_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b32_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelt.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b32,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelt.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b64_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelt.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b64_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b64_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelt.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b64,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b8_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelo.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilelt_b8_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelo.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilelt_b8_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b8_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelo.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilelt_b8,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelo.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b16_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelo.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b16_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b16_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelo.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b16,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelo.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b32_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelo.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b32_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b32_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelo.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b32,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelo.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b64_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelo.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b64_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b64_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelo.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b64,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b8_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelt.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilelt_b8_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelt.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilelt_b8_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b8_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelt.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilelt_b8,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelt.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b16_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelt.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b16_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b16_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelt.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b16,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelt.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b32_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelt.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b32_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b32_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelt.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b32,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelt.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b64_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelt.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b64_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b64_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelt.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b64,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b8_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelo.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilelt_b8_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelo.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilelt_b8_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b8_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilelo.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svwhilelt_b8,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelo.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b16_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelo.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b16_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b16_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilelo.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b16,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelo.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b32_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelo.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b32_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b32_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilelo.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b32,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilelt_b64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelo.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilelt_b64_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelo.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilelt_b64_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilelt_b64_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilelo.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return SVE_ACLE_FUNC(svwhilelt_b64,_u64,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_wrffr.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_wrffr.c
index ea9ae5ddb1dc8..a1e3ae0f0d9cb 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_wrffr.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_wrffr.c
@@ -1,13 +1,21 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o - %s >/dev/null
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -o /dev/null %s
 #include <arm_sve.h>
 
+// CHECK-LABEL: @test_svwrffr(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    call void @llvm.aarch64.sve.wrffr(<vscale x 16 x i1> [[OP:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z12test_svwrffru10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.wrffr(<vscale x 16 x i1> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svwrffr(svbool_t op)
 {
-  // CHECK-LABEL: test_svwrffr
-  // CHECK: call void @llvm.aarch64.sve.wrffr(<vscale x 16 x i1> %op)
-  // CHECK: ret void
   svwrffr(op);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-bfloat.c
index 7bcf840e542ce..cec3a2af17083 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip1.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svzip1_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip1.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svzip1_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svzip1_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip1.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svzip1_bf16'}}
   return SVE_ACLE_FUNC(svzip1,_bf16,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64-bfloat.c
index 218e8413454af..17a9407ed09ed 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip1_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip1q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svzip1_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip1q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svzip1_bf16(svbfloat16_t op1, svbfloat16_t op2) {
-  // CHECK-LABEL: test_svzip1_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip1q.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svzip1q_bf16'}}
   return SVE_ACLE_FUNC(svzip1q, _bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64.c
index 88ec67c309e18..49da35c0ceddb 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1-fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,79 +13,156 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip1_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svzip1_s8(svint8_t op1, svint8_t op2) {
-  // CHECK-LABEL: test_svzip1_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _s8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svzip1_s16(svint16_t op1, svint16_t op2) {
-  // CHECK-LABEL: test_svzip1_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _s16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svzip1_s32(svint32_t op1, svint32_t op2) {
-  // CHECK-LABEL: test_svzip1_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _s32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svzip1_s64(svint64_t op1, svint64_t op2) {
-  // CHECK-LABEL: test_svzip1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _s64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip1_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svzip1_u8(svuint8_t op1, svuint8_t op2) {
-  // CHECK-LABEL: test_svzip1_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _u8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svzip1_u16(svuint16_t op1, svuint16_t op2) {
-  // CHECK-LABEL: test_svzip1_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _u16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svzip1_u32(svuint32_t op1, svuint32_t op2) {
-  // CHECK-LABEL: test_svzip1_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _u32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svzip1_u64(svuint64_t op1, svuint64_t op2) {
-  // CHECK-LABEL: test_svzip1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _u64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip1q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip1q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svzip1_f16(svfloat16_t op1, svfloat16_t op2) {
-  // CHECK-LABEL: test_svzip1_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip1q.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _f16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip1q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip1q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svzip1_f32(svfloat32_t op1, svfloat32_t op2) {
-  // CHECK-LABEL: test_svzip1_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip1q.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _f32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip1q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip1q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svzip1_f64(svfloat64_t op1, svfloat64_t op2) {
-  // CHECK-LABEL: test_svzip1_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip1q.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1q, _f64, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1.c
index dbd6489945358..ca98f6fdce1e7 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip1.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,131 +14,245 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip1_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip1_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svzip1_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svzip1_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svzip1_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svzip1_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svzip1_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svzip1_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svzip1_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svzip1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip1_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svzip1_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svzip1_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip1.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svzip1_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svzip1_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip1.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svzip1_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svzip1_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip1.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svzip1_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svzip1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip1.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip1.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip1.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svzip1_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svzip1_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip1.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip1.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip1.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svzip1_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svzip1_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip1.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip1.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip1.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svzip1_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svzip1_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip1.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip1,_f64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip1.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip1_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip1.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svzip1_b8(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip1_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip1.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svzip1_b8(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip1.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip1.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svzip1_b16(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip1_b16
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip1.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svzip1_b16(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip1.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip1.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svzip1_b32(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip1_b32
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip1.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svzip1_b32(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip1_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip1.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip1_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip1.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svzip1_b64(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip1_b64
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip1.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svzip1_b64(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-bfloat.c
index c86a0e518c5bd..4f72d855f52af 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,11 +14,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip2.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svzip2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip2.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svzip2_bf16(svbfloat16_t op1, svbfloat16_t op2)
 {
-  // CHECK-LABEL: test_svzip2_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip2.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svzip2_bf16'}}
   return SVE_ACLE_FUNC(svzip2,_bf16,,)(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64-bfloat.c
index 626324f0e32ae..531e2728bb3ba 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,10 +14,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip2q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svzip2_bf16u14__SVBFloat16_tu14__SVBFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip2q.nxv8bf16(<vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svzip2_bf16(svbfloat16_t op1, svbfloat16_t op2) {
-  // CHECK-LABEL: test_svzip2_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zip2q.nxv8bf16(<vscale x 8 x bfloat> %op1, <vscale x 8 x bfloat> %op2)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svzip2q_bf16'}}
   return SVE_ACLE_FUNC(svzip2q, _bf16, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64.c
index f4ff36c2ee6be..c10c375c31efe 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2-fp64.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -target-feature +f64mm -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 
 #include <arm_sve.h>
 
@@ -12,79 +13,156 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svzip2_s8(svint8_t op1, svint8_t op2) {
-  // CHECK-LABEL: test_svzip2_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _s8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svzip2_s16(svint16_t op1, svint16_t op2) {
-  // CHECK-LABEL: test_svzip2_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _s16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svzip2_s32(svint32_t op1, svint32_t op2) {
-  // CHECK-LABEL: test_svzip2_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _s32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svzip2_s64(svint64_t op1, svint64_t op2) {
-  // CHECK-LABEL: test_svzip2_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _s64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2q.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svzip2_u8(svuint8_t op1, svuint8_t op2) {
-  // CHECK-LABEL: test_svzip2_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2q.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _u8, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2q.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svzip2_u16(svuint16_t op1, svuint16_t op2) {
-  // CHECK-LABEL: test_svzip2_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2q.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _u16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2q.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svzip2_u32(svuint32_t op1, svuint32_t op2) {
-  // CHECK-LABEL: test_svzip2_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2q.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _u32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2q.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svzip2_u64(svuint64_t op1, svuint64_t op2) {
-  // CHECK-LABEL: test_svzip2_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2q.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _u64, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2q.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svzip2_f16(svfloat16_t op1, svfloat16_t op2) {
-  // CHECK-LABEL: test_svzip2_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2q.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _f16, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2q.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svzip2_f32(svfloat32_t op1, svfloat32_t op2) {
-  // CHECK-LABEL: test_svzip2_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2q.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _f32, , )(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2q.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svzip2_f64(svfloat64_t op1, svfloat64_t op2) {
-  // CHECK-LABEL: test_svzip2_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2q.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2q, _f64, , )(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2.c
index 839429246d65b..25676f9f10a9e 100644
--- a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2.c
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_zip2.c
@@ -1,9 +1,10 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o - %s >/dev/null
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -o /dev/null %s
 #include <arm_sve.h>
 
 #ifdef SVE_OVERLOADED_FORMS
@@ -13,131 +14,245 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svzip2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip2_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svzip2_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svzip2_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svzip2_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svzip2_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svzip2_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svzip2_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svzip2_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svzip2_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip2_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svzip2_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svzip2_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.zip2.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svzip2_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svzip2_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.zip2.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svzip2_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svzip2_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.zip2.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svzip2_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svzip2_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.zip2.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_f16u13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2.nxv8f16(<vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svzip2_f16(svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svzip2_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.zip2.nxv8f16(<vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_f32u13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svzip2_f32(svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svzip2_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.zip2.nxv4f32(<vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_f64u13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2.nxv2f64(<vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svzip2_f64(svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svzip2_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.zip2.nxv2f64(<vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   return SVE_ACLE_FUNC(svzip2,_f64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_b8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svzip2_b8u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip2.nxv16i1(<vscale x 16 x i1> [[OP1:%.*]], <vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svzip2_b8(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip2_b8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.zip2.nxv16i1(<vscale x 16 x i1> %op1, <vscale x 16 x i1> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   return svzip2_b8(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_b16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip2.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_b16u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip2.nxv8i1(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svzip2_b16(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip2_b16
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.zip2.nxv8i1(<vscale x 8 x i1> %[[OP1]], <vscale x 8 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svzip2_b16(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_b32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip2.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_b32u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip2.nxv4i1(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svzip2_b32(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip2_b32
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.zip2.nxv4i1(<vscale x 4 x i1> %[[OP1]], <vscale x 4 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svzip2_b32(op1, op2);
 }
 
+// CHECK-LABEL: @test_svzip2_b64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip2.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svzip2_b64u10__SVBool_tu10__SVBool_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip2.nxv2i1(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP2]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP3]]
+//
 svbool_t test_svzip2_b64(svbool_t op1, svbool_t op2)
 {
-  // CHECK-LABEL: test_svzip2_b64
-  // CHECK-DAG: %[[OP1:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op1)
-  // CHECK-DAG: %[[OP2:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.zip2.nxv2i1(<vscale x 2 x i1> %[[OP1]], <vscale x 2 x i1> %[[OP2]])
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   return svzip2_b64(op1, op2);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aba.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aba.c
index 57ef110e09204..64b2a14346525 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aba.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aba.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaba_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.saba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svaba_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.saba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svaba_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svaba_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.saba.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_s8'}}
   return SVE_ACLE_FUNC(svaba,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaba_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaba_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svaba_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saba.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_s16'}}
   return SVE_ACLE_FUNC(svaba,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaba_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaba_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svaba_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saba.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_s32'}}
   return SVE_ACLE_FUNC(svaba,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaba_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svaba_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svaba_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saba.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_s64'}}
   return SVE_ACLE_FUNC(svaba,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uaba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svaba_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uaba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaba_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svaba_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uaba.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_u8'}}
   return SVE_ACLE_FUNC(svaba,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaba_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaba_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svaba_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaba.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_u16'}}
   return SVE_ACLE_FUNC(svaba,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaba_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaba_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svaba_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaba.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_u32'}}
   return SVE_ACLE_FUNC(svaba,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaba_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svaba_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svaba_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaba.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_u64'}}
   return SVE_ACLE_FUNC(svaba,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.saba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaba_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.saba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svaba_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.saba.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_s8'}}
   return SVE_ACLE_FUNC(svaba,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaba_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaba_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saba.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_s16'}}
   return SVE_ACLE_FUNC(svaba,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaba_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaba_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saba.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_s32'}}
   return SVE_ACLE_FUNC(svaba,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaba_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaba_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saba.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_s64'}}
   return SVE_ACLE_FUNC(svaba,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uaba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaba_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uaba.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svaba_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uaba.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_u8'}}
   return SVE_ACLE_FUNC(svaba,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaba_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaba.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaba_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaba.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_u16'}}
   return SVE_ACLE_FUNC(svaba,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaba_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaba.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaba_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaba.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_u32'}}
   return SVE_ACLE_FUNC(svaba,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaba_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaba_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaba.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaba_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svaba_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaba.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaba'}}
   // expected-warning at +1 {{implicit declaration of function 'svaba_n_u64'}}
   return SVE_ACLE_FUNC(svaba,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalb.c
index 28f4d38f81f11..fcb0eedd6be46 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svabalb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalb_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svabalb_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svabalb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_s16'}}
   return SVE_ACLE_FUNC(svabalb,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalb_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svabalb_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svabalb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_s32'}}
   return SVE_ACLE_FUNC(svabalb,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalb_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svabalb_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svabalb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_s64'}}
   return SVE_ACLE_FUNC(svabalb,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalb_u16u12__SVUint16_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svabalb_u16(svuint16_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svabalb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_u16'}}
   return SVE_ACLE_FUNC(svabalb,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalb_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svabalb_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svabalb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_u32'}}
   return SVE_ACLE_FUNC(svabalb,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalb_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svabalb_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svabalb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_u64'}}
   return SVE_ACLE_FUNC(svabalb,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalb_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabalb_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svabalb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_n_s16'}}
   return SVE_ACLE_FUNC(svabalb,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalb_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabalb_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svabalb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_n_s32'}}
   return SVE_ACLE_FUNC(svabalb,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalb_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabalb_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svabalb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_n_s64'}}
   return SVE_ACLE_FUNC(svabalb,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalb_n_u16u12__SVUint16_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svabalb_n_u16(svuint16_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svabalb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_n_u16'}}
   return SVE_ACLE_FUNC(svabalb,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalb_n_u32u12__SVUint32_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svabalb_n_u32(svuint32_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svabalb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_n_u32'}}
   return SVE_ACLE_FUNC(svabalb,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalb_n_u64u12__SVUint64_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svabalb_n_u64(svuint64_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svabalb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalb_n_u64'}}
   return SVE_ACLE_FUNC(svabalb,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalt.c
index fb281da93b0df..246f84b2273ba 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abalt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svabalt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svabalt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svabalt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_s16'}}
   return SVE_ACLE_FUNC(svabalt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svabalt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svabalt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_s32'}}
   return SVE_ACLE_FUNC(svabalt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svabalt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svabalt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_s64'}}
   return SVE_ACLE_FUNC(svabalt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalt_u16u12__SVUint16_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svabalt_u16(svuint16_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svabalt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_u16'}}
   return SVE_ACLE_FUNC(svabalt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalt_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svabalt_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svabalt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_u32'}}
   return SVE_ACLE_FUNC(svabalt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabalt_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svabalt_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svabalt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_u64'}}
   return SVE_ACLE_FUNC(svabalt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabalt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svabalt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_n_s16'}}
   return SVE_ACLE_FUNC(svabalt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabalt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svabalt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_n_s32'}}
   return SVE_ACLE_FUNC(svabalt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabalt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svabalt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_n_s64'}}
   return SVE_ACLE_FUNC(svabalt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalt_n_u16u12__SVUint16_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svabalt_n_u16(svuint16_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svabalt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_n_u16'}}
   return SVE_ACLE_FUNC(svabalt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalt_n_u32u12__SVUint32_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svabalt_n_u32(svuint32_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svabalt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_n_u32'}}
   return SVE_ACLE_FUNC(svabalt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svabalt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabalt_n_u64u12__SVUint64_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svabalt_n_u64(svuint64_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svabalt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabalt_n_u64'}}
   return SVE_ACLE_FUNC(svabalt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlb.c
index 1c59ec6110a5a..d91398c81deca 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svabdlb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlb_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svabdlb_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_s16'}}
   return SVE_ACLE_FUNC(svabdlb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlb_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svabdlb_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_s32'}}
   return SVE_ACLE_FUNC(svabdlb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlb_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svabdlb_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_s64'}}
   return SVE_ACLE_FUNC(svabdlb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlb_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svabdlb_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_u16'}}
   return SVE_ACLE_FUNC(svabdlb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlb_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svabdlb_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_u32'}}
   return SVE_ACLE_FUNC(svabdlb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlb_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svabdlb_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_u64'}}
   return SVE_ACLE_FUNC(svabdlb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlb_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabdlb_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_n_s16'}}
   return SVE_ACLE_FUNC(svabdlb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlb_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabdlb_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_n_s32'}}
   return SVE_ACLE_FUNC(svabdlb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlb_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabdlb_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_n_s64'}}
   return SVE_ACLE_FUNC(svabdlb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlb_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svabdlb_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_n_u16'}}
   return SVE_ACLE_FUNC(svabdlb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlb_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svabdlb_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_n_u32'}}
   return SVE_ACLE_FUNC(svabdlb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlb_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svabdlb_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svabdlb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlb_n_u64'}}
   return SVE_ACLE_FUNC(svabdlb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlt.c
index ed5a692d94a24..491ac415dde25 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_abdlt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svabdlt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svabdlt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_s16'}}
   return SVE_ACLE_FUNC(svabdlt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svabdlt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_s32'}}
   return SVE_ACLE_FUNC(svabdlt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svabdlt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_s64'}}
   return SVE_ACLE_FUNC(svabdlt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlt_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svabdlt_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_u16'}}
   return SVE_ACLE_FUNC(svabdlt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlt_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svabdlt_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_u32'}}
   return SVE_ACLE_FUNC(svabdlt,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svabdlt_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svabdlt_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_u64'}}
   return SVE_ACLE_FUNC(svabdlt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svabdlt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sabdlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_n_s16'}}
   return SVE_ACLE_FUNC(svabdlt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svabdlt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sabdlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_n_s32'}}
   return SVE_ACLE_FUNC(svabdlt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svabdlt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sabdlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_n_s64'}}
   return SVE_ACLE_FUNC(svabdlt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlt_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svabdlt_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uabdlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_n_u16'}}
   return SVE_ACLE_FUNC(svabdlt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlt_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svabdlt_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uabdlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_n_u32'}}
   return SVE_ACLE_FUNC(svabdlt,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svabdlt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svabdlt_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svabdlt_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svabdlt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uabdlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svabdlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svabdlt_n_u64'}}
   return SVE_ACLE_FUNC(svabdlt,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adalp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adalp.c
index 31f4d4a74347c..123880b715917 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adalp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adalp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,205 +15,355 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadalp_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s16_zu10__SVBool_tu11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svadalp_s16_z(svbool_t pg, svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s16_z'}}
   return SVE_ACLE_FUNC(svadalp,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svadalp_s32_z(svbool_t pg, svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s32_z'}}
   return SVE_ACLE_FUNC(svadalp,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svadalp_s64_z(svbool_t pg, svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s64_z'}}
   return SVE_ACLE_FUNC(svadalp,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u16_zu10__SVBool_tu12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svadalp_u16_z(svbool_t pg, svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u16_z'}}
   return SVE_ACLE_FUNC(svadalp,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svadalp_u32_z(svbool_t pg, svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u32_z'}}
   return SVE_ACLE_FUNC(svadalp,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svadalp_u64_z(svbool_t pg, svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u64_z'}}
   return SVE_ACLE_FUNC(svadalp,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s16_mu10__SVBool_tu11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svadalp_s16_m(svbool_t pg, svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s16_m'}}
   return SVE_ACLE_FUNC(svadalp,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svadalp_s32_m(svbool_t pg, svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s32_m'}}
   return SVE_ACLE_FUNC(svadalp,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svadalp_s64_m(svbool_t pg, svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s64_m'}}
   return SVE_ACLE_FUNC(svadalp,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u16_mu10__SVBool_tu12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svadalp_u16_m(svbool_t pg, svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u16_m'}}
   return SVE_ACLE_FUNC(svadalp,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svadalp_u32_m(svbool_t pg, svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u32_m'}}
   return SVE_ACLE_FUNC(svadalp,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svadalp_u64_m(svbool_t pg, svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u64_m'}}
   return SVE_ACLE_FUNC(svadalp,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s16_xu10__SVBool_tu11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svadalp_s16_x(svbool_t pg, svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sadalp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s16_x'}}
   return SVE_ACLE_FUNC(svadalp,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svadalp_s32_x(svbool_t pg, svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sadalp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s32_x'}}
   return SVE_ACLE_FUNC(svadalp,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svadalp_s64_x(svbool_t pg, svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svadalp_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sadalp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_s64_x'}}
   return SVE_ACLE_FUNC(svadalp,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u16_xu10__SVBool_tu12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svadalp_u16_x(svbool_t pg, svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uadalp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u16_x'}}
   return SVE_ACLE_FUNC(svadalp,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svadalp_u32_x(svbool_t pg, svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uadalp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u32_x'}}
   return SVE_ACLE_FUNC(svadalp,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svadalp_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadalp_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svadalp_u64_x(svbool_t pg, svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svadalp_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uadalp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadalp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svadalp_u64_x'}}
   return SVE_ACLE_FUNC(svadalp,_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclb.c
index 55c0f3c70577b..dc9bfbc58e503 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,43 +15,73 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadclb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadclb_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadclb_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svadclb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclb_u32'}}
   return SVE_ACLE_FUNC(svadclb,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svadclb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadclb_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadclb_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svadclb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclb_u64'}}
   return SVE_ACLE_FUNC(svadclb,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svadclb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadclb_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svadclb_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svadclb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclb_n_u32'}}
   return SVE_ACLE_FUNC(svadclb,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svadclb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadclb_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svadclb_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svadclb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclb_n_u64'}}
   return SVE_ACLE_FUNC(svadclb,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclt.c
index 4f11e63fd41ea..a42ac072e13e2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_adclt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,43 +15,73 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svadclt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadclt_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svadclt_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svadclt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclt_u32'}}
   return SVE_ACLE_FUNC(svadclt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svadclt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svadclt_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svadclt_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svadclt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclt_u64'}}
   return SVE_ACLE_FUNC(svadclt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svadclt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadclt_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svadclt_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svadclt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.adclt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclt_n_u32'}}
   return SVE_ACLE_FUNC(svadclt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svadclt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svadclt_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svadclt_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svadclt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.adclt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svadclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svadclt_n_u64'}}
   return SVE_ACLE_FUNC(svadclt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnb.c
index e325d0098caa0..070fc493b9ee4 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddhnb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnb_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svaddhnb_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_s16'}}
   return SVE_ACLE_FUNC(svaddhnb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnb_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddhnb_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_s32'}}
   return SVE_ACLE_FUNC(svaddhnb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnb_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddhnb_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_s64'}}
   return SVE_ACLE_FUNC(svaddhnb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnb_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaddhnb_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_u16'}}
   return SVE_ACLE_FUNC(svaddhnb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnb_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaddhnb_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_u32'}}
   return SVE_ACLE_FUNC(svaddhnb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnb_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaddhnb_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_u64'}}
   return SVE_ACLE_FUNC(svaddhnb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnb_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svaddhnb_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_n_s16'}}
   return SVE_ACLE_FUNC(svaddhnb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnb_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddhnb_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_n_s32'}}
   return SVE_ACLE_FUNC(svaddhnb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnb_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddhnb_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_n_s64'}}
   return SVE_ACLE_FUNC(svaddhnb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnb_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svaddhnb_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_n_u16'}}
   return SVE_ACLE_FUNC(svaddhnb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnb_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddhnb_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_n_u32'}}
   return SVE_ACLE_FUNC(svaddhnb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddhnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnb_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddhnb_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svaddhnb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnb_n_u64'}}
   return SVE_ACLE_FUNC(svaddhnb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnt.c
index bcb68970513b7..8bc757ec9356a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addhnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddhnt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnt_s16u10__SVInt8_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svaddhnt_s16(svint8_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_s16'}}
   return SVE_ACLE_FUNC(svaddhnt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnt_s32u11__SVInt16_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddhnt_s32(svint16_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_s32'}}
   return SVE_ACLE_FUNC(svaddhnt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnt_s64u11__SVInt32_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddhnt_s64(svint32_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_s64'}}
   return SVE_ACLE_FUNC(svaddhnt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnt_u16u11__SVUint8_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaddhnt_u16(svuint8_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_u16'}}
   return SVE_ACLE_FUNC(svaddhnt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnt_u32u12__SVUint16_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaddhnt_u32(svuint16_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_u32'}}
   return SVE_ACLE_FUNC(svaddhnt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddhnt_u64u12__SVUint32_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaddhnt_u64(svuint32_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_u64'}}
   return SVE_ACLE_FUNC(svaddhnt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnt_n_s16u10__SVInt8_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svaddhnt_n_s16(svint8_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_n_s16'}}
   return SVE_ACLE_FUNC(svaddhnt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnt_n_s32u11__SVInt16_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddhnt_n_s32(svint16_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_n_s32'}}
   return SVE_ACLE_FUNC(svaddhnt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnt_n_s64u11__SVInt32_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddhnt_n_s64(svint32_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_n_s64'}}
   return SVE_ACLE_FUNC(svaddhnt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnt_n_u16u11__SVUint8_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svaddhnt_n_u16(svuint8_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_n_u16'}}
   return SVE_ACLE_FUNC(svaddhnt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnt_n_u32u12__SVUint16_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddhnt_n_u32(svuint16_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_n_u32'}}
   return SVE_ACLE_FUNC(svaddhnt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svaddhnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddhnt_n_u64u12__SVUint32_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddhnt_n_u64(svuint32_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svaddhnt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddhnt_n_u64'}}
   return SVE_ACLE_FUNC(svaddhnt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlb.c
index affe1c78393ed..8eed03d66bee2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddlb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlb_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddlb_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_s16'}}
   return SVE_ACLE_FUNC(svaddlb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlb_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddlb_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_s32'}}
   return SVE_ACLE_FUNC(svaddlb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlb_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svaddlb_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_s64'}}
   return SVE_ACLE_FUNC(svaddlb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlb_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaddlb_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_u16'}}
   return SVE_ACLE_FUNC(svaddlb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlb_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaddlb_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_u32'}}
   return SVE_ACLE_FUNC(svaddlb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlb_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svaddlb_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_u64'}}
   return SVE_ACLE_FUNC(svaddlb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlb_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddlb_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_n_s16'}}
   return SVE_ACLE_FUNC(svaddlb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlb_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddlb_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_n_s32'}}
   return SVE_ACLE_FUNC(svaddlb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlb_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddlb_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_n_s64'}}
   return SVE_ACLE_FUNC(svaddlb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlb_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddlb_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_n_u16'}}
   return SVE_ACLE_FUNC(svaddlb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlb_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddlb_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_n_u32'}}
   return SVE_ACLE_FUNC(svaddlb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlb_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaddlb_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlb_n_u64'}}
   return SVE_ACLE_FUNC(svaddlb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlbt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlbt.c
index c729ea91675f3..adbc91e186691 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlbt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlbt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,64 +15,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddlbt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddlbt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddlbt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlbt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlbt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlbt_s16'}}
   return SVE_ACLE_FUNC(svaddlbt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlbt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddlbt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddlbt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlbt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlbt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlbt_s32'}}
   return SVE_ACLE_FUNC(svaddlbt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlbt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddlbt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svaddlbt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlbt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlbt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlbt_s64'}}
   return SVE_ACLE_FUNC(svaddlbt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlbt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddlbt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddlbt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svaddlbt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlbt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlbt_n_s16'}}
   return SVE_ACLE_FUNC(svaddlbt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlbt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddlbt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddlbt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svaddlbt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlbt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlbt_n_s32'}}
   return SVE_ACLE_FUNC(svaddlbt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlbt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svaddlbt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddlbt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svaddlbt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlbt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlbt_n_s64'}}
   return SVE_ACLE_FUNC(svaddlbt,_n_s64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlt.c
index 88a0966e897e9..80bbcec99e8cb 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addlt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddlt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddlt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_s16'}}
   return SVE_ACLE_FUNC(svaddlt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddlt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_s32'}}
   return SVE_ACLE_FUNC(svaddlt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svaddlt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_s64'}}
   return SVE_ACLE_FUNC(svaddlt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlt_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaddlt_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_u16'}}
   return SVE_ACLE_FUNC(svaddlt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlt_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaddlt_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_u32'}}
   return SVE_ACLE_FUNC(svaddlt,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddlt_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svaddlt_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_u64'}}
   return SVE_ACLE_FUNC(svaddlt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddlt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_n_s16'}}
   return SVE_ACLE_FUNC(svaddlt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddlt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_n_s32'}}
   return SVE_ACLE_FUNC(svaddlt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddlt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_n_s64'}}
   return SVE_ACLE_FUNC(svaddlt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlt_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddlt_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddlt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_n_u16'}}
   return SVE_ACLE_FUNC(svaddlt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlt_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddlt_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddlt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_n_u32'}}
   return SVE_ACLE_FUNC(svaddlt,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddlt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddlt_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaddlt_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svaddlt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddlt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddlt_n_u64'}}
   return SVE_ACLE_FUNC(svaddlt,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addp.c
index 0cf04ccf681ee..781d929b11f9d 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,239 +15,411 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddp_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddp_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svaddp_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s8_m'}}
   return SVE_ACLE_FUNC(svaddp,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddp_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s16_m'}}
   return SVE_ACLE_FUNC(svaddp,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddp_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s32_m'}}
   return SVE_ACLE_FUNC(svaddp,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddp_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s64_m'}}
   return SVE_ACLE_FUNC(svaddp,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddp_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaddp_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u8_m'}}
   return SVE_ACLE_FUNC(svaddp,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddp_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u16_m'}}
   return SVE_ACLE_FUNC(svaddp,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddp_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u32_m'}}
   return SVE_ACLE_FUNC(svaddp,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaddp_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u64_m'}}
   return SVE_ACLE_FUNC(svaddp,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddp_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svaddp_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s8_x'}}
   return SVE_ACLE_FUNC(svaddp,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddp_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s16_x'}}
   return SVE_ACLE_FUNC(svaddp,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddp_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s32_x'}}
   return SVE_ACLE_FUNC(svaddp,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddp_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svaddp_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_s64_x'}}
   return SVE_ACLE_FUNC(svaddp,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddp_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaddp_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.addp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u8_x'}}
   return SVE_ACLE_FUNC(svaddp,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddp_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.addp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u16_x'}}
   return SVE_ACLE_FUNC(svaddp,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddp_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.addp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u32_x'}}
   return SVE_ACLE_FUNC(svaddp,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaddp_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svaddp_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.addp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_u64_x'}}
   return SVE_ACLE_FUNC(svaddp,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.faddp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.faddp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svaddp_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svaddp_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.faddp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_f16_m'}}
   return SVE_ACLE_FUNC(svaddp,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.faddp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.faddp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svaddp_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svaddp_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.faddp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_f32_m'}}
   return SVE_ACLE_FUNC(svaddp,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.faddp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.faddp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svaddp_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svaddp_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.faddp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_f64_m'}}
   return SVE_ACLE_FUNC(svaddp,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.faddp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.faddp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svaddp_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svaddp_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.faddp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_f16_x'}}
   return SVE_ACLE_FUNC(svaddp,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.faddp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.faddp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svaddp_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svaddp_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.faddp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_f32_x'}}
   return SVE_ACLE_FUNC(svaddp,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddp_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.faddp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svaddp_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.faddp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svaddp_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svaddp_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.faddp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddp_f64_x'}}
   return SVE_ACLE_FUNC(svaddp,_f64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwb.c
index cd3519c3217ff..dd0fecaf0fcd0 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddwb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwb_s16u11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddwb_s16(svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_s16'}}
   return SVE_ACLE_FUNC(svaddwb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwb_s32u11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddwb_s32(svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_s32'}}
   return SVE_ACLE_FUNC(svaddwb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwb_s64u11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svaddwb_s64(svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_s64'}}
   return SVE_ACLE_FUNC(svaddwb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwb_u16u12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaddwb_u16(svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_u16'}}
   return SVE_ACLE_FUNC(svaddwb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwb_u32u12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaddwb_u32(svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_u32'}}
   return SVE_ACLE_FUNC(svaddwb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwb_u64u12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svaddwb_u64(svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_u64'}}
   return SVE_ACLE_FUNC(svaddwb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwb_n_s16u11__SVInt16_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddwb_n_s16(svint16_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_n_s16'}}
   return SVE_ACLE_FUNC(svaddwb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwb_n_s32u11__SVInt32_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddwb_n_s32(svint32_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_n_s32'}}
   return SVE_ACLE_FUNC(svaddwb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwb_n_s64u11__SVInt64_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddwb_n_s64(svint64_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_n_s64'}}
   return SVE_ACLE_FUNC(svaddwb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwb_n_u16u12__SVUint16_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddwb_n_u16(svuint16_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_n_u16'}}
   return SVE_ACLE_FUNC(svaddwb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwb_n_u32u12__SVUint32_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddwb_n_u32(svuint32_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_n_u32'}}
   return SVE_ACLE_FUNC(svaddwb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwb_n_u64u12__SVUint64_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaddwb_n_u64(svuint64_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svaddwb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwb_n_u64'}}
   return SVE_ACLE_FUNC(svaddwb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwt.c
index d038c42d9d0a3..b059f5c698ab4 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_addwt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaddwt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwt_s16u11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svaddwt_s16(svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_s16'}}
   return SVE_ACLE_FUNC(svaddwt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwt_s32u11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svaddwt_s32(svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_s32'}}
   return SVE_ACLE_FUNC(svaddwt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwt_s64u11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svaddwt_s64(svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_s64'}}
   return SVE_ACLE_FUNC(svaddwt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwt_u16u12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svaddwt_u16(svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_u16'}}
   return SVE_ACLE_FUNC(svaddwt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwt_u32u12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svaddwt_u32(svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_u32'}}
   return SVE_ACLE_FUNC(svaddwt,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaddwt_u64u12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svaddwt_u64(svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_u64'}}
   return SVE_ACLE_FUNC(svaddwt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwt_n_s16u11__SVInt16_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svaddwt_n_s16(svint16_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.saddwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_n_s16'}}
   return SVE_ACLE_FUNC(svaddwt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwt_n_s32u11__SVInt32_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svaddwt_n_s32(svint32_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.saddwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_n_s32'}}
   return SVE_ACLE_FUNC(svaddwt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwt_n_s64u11__SVInt64_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svaddwt_n_s64(svint64_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.saddwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_n_s64'}}
   return SVE_ACLE_FUNC(svaddwt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwt_n_u16u12__SVUint16_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svaddwt_n_u16(svuint16_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uaddwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_n_u16'}}
   return SVE_ACLE_FUNC(svaddwt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwt_n_u32u12__SVUint32_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svaddwt_n_u32(svuint32_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uaddwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_n_u32'}}
   return SVE_ACLE_FUNC(svaddwt,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svaddwt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svaddwt_n_u64u12__SVUint64_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svaddwt_n_u64(svuint64_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svaddwt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uaddwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaddwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svaddwt_n_u64'}}
   return SVE_ACLE_FUNC(svaddwt,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesd.c
index e9290948268fd..3f8cd967eb514 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,11 +15,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaesd_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesd(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaesd_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesd(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaesd_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaesd_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesd(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaesd'}}
   // expected-warning at +1 {{implicit declaration of function 'svaesd_u8'}}
   return SVE_ACLE_FUNC(svaesd,_u8,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aese.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aese.c
index c10060a8aef5e..9256ab70cab99 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aese.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aese.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,11 +15,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaese_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aese(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svaese_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aese(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaese_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svaese_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aese(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaese'}}
   // expected-warning at +1 {{implicit declaration of function 'svaese_u8'}}
   return SVE_ACLE_FUNC(svaese,_u8,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesimc.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesimc.c
index d4ed607012c0b..627d22f45b683 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesimc.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesimc.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,11 +15,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaesimc_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesimc(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svaesimc_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesimc(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaesimc_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svaesimc_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesimc(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaesimc'}}
   // expected-warning at +1 {{implicit declaration of function 'svaesimc_u8'}}
   return SVE_ACLE_FUNC(svaesimc,_u8,,)(op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesmc.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesmc.c
index 25bd65d38a4fa..1b96d9aafa043 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesmc.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_aesmc.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,11 +15,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svaesmc_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesmc(<vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svaesmc_u8u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesmc(<vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svaesmc_u8(svuint8_t op)
 {
-  // CHECK-LABEL: test_svaesmc_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.aesmc(<vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svaesmc'}}
   // expected-warning at +1 {{implicit declaration of function 'svaesmc_u8'}}
   return SVE_ACLE_FUNC(svaesmc,_u8,,)(op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bcax.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bcax.c
index 874f111bf09f1..7d68aef34a446 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bcax.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bcax.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbcax_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbcax_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svbcax_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svbcax_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_s8'}}
   return SVE_ACLE_FUNC(svbcax,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbcax_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svbcax_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svbcax_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_s16'}}
   return SVE_ACLE_FUNC(svbcax,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbcax_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svbcax_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svbcax_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_s32'}}
   return SVE_ACLE_FUNC(svbcax,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbcax_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svbcax_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svbcax_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_s64'}}
   return SVE_ACLE_FUNC(svbcax,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbcax_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbcax_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svbcax_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_u8'}}
   return SVE_ACLE_FUNC(svbcax,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbcax_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbcax_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svbcax_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_u16'}}
   return SVE_ACLE_FUNC(svbcax,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbcax_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbcax_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svbcax_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_u32'}}
   return SVE_ACLE_FUNC(svbcax,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbcax_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbcax_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svbcax_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_u64'}}
   return SVE_ACLE_FUNC(svbcax,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbcax_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbcax_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_s8'}}
   return SVE_ACLE_FUNC(svbcax,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbcax_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svbcax_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_s16'}}
   return SVE_ACLE_FUNC(svbcax,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbcax_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svbcax_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_s32'}}
   return SVE_ACLE_FUNC(svbcax,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbcax_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svbcax_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_s64'}}
   return SVE_ACLE_FUNC(svbcax,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbcax_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbcax_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_u8'}}
   return SVE_ACLE_FUNC(svbcax,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbcax_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbcax_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_u16'}}
   return SVE_ACLE_FUNC(svbcax,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbcax_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbcax_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_u32'}}
   return SVE_ACLE_FUNC(svbcax,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbcax_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbcax_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbcax_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svbcax_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbcax'}}
   // expected-warning at +1 {{implicit declaration of function 'svbcax_n_u64'}}
   return SVE_ACLE_FUNC(svbcax,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bdep.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bdep.c
index f30c16dd2a742..feb02c035e30a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bdep.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bdep.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,85 +15,145 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbdep_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bdep.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbdep_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bdep.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbdep_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svbdep_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bdep.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_u8'}}
   return SVE_ACLE_FUNC(svbdep,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bdep.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbdep_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bdep.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbdep_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svbdep_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bdep.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_u16'}}
   return SVE_ACLE_FUNC(svbdep,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bdep.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbdep_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bdep.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbdep_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svbdep_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bdep.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_u32'}}
   return SVE_ACLE_FUNC(svbdep,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bdep.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbdep_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bdep.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbdep_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svbdep_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bdep.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_u64'}}
   return SVE_ACLE_FUNC(svbdep,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bdep.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbdep_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bdep.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbdep_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svbdep_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bdep.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_n_u8'}}
   return SVE_ACLE_FUNC(svbdep,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bdep.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbdep_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bdep.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbdep_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svbdep_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bdep.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_n_u16'}}
   return SVE_ACLE_FUNC(svbdep,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bdep.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbdep_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bdep.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbdep_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svbdep_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bdep.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_n_u32'}}
   return SVE_ACLE_FUNC(svbdep,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbdep_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bdep.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbdep_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bdep.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbdep_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svbdep_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bdep.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbdep'}}
   // expected-warning at +1 {{implicit declaration of function 'svbdep_n_u64'}}
   return SVE_ACLE_FUNC(svbdep,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bext.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bext.c
index 462c87044bc24..2e351c65b83a0 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bext.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bext.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,85 +15,145 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbext_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bext.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbext_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bext.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbext_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svbext_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bext.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_u8'}}
   return SVE_ACLE_FUNC(svbext,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bext.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbext_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bext.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbext_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svbext_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bext.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_u16'}}
   return SVE_ACLE_FUNC(svbext,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bext.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbext_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bext.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbext_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svbext_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bext.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_u32'}}
   return SVE_ACLE_FUNC(svbext,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bext.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbext_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bext.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbext_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svbext_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bext.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_u64'}}
   return SVE_ACLE_FUNC(svbext,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bext.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbext_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bext.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbext_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svbext_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bext.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_n_u8'}}
   return SVE_ACLE_FUNC(svbext,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bext.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbext_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bext.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbext_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svbext_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bext.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_n_u16'}}
   return SVE_ACLE_FUNC(svbext,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bext.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbext_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bext.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbext_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svbext_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bext.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_n_u32'}}
   return SVE_ACLE_FUNC(svbext,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbext_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bext.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbext_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bext.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbext_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svbext_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bext.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbext'}}
   // expected-warning at +1 {{implicit declaration of function 'svbext_n_u64'}}
   return SVE_ACLE_FUNC(svbext,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bgrp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bgrp.c
index b8b178dbde8a7..88d2abb048081 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bgrp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bgrp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-bitperm -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,85 +15,145 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbgrp_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bgrp.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbgrp_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bgrp.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbgrp_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bgrp.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_u8'}}
   return SVE_ACLE_FUNC(svbgrp,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bgrp.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbgrp_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bgrp.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbgrp_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bgrp.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_u16'}}
   return SVE_ACLE_FUNC(svbgrp,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bgrp.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbgrp_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bgrp.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbgrp_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bgrp.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_u32'}}
   return SVE_ACLE_FUNC(svbgrp,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bgrp.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbgrp_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bgrp.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbgrp_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bgrp.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_u64'}}
   return SVE_ACLE_FUNC(svbgrp,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bgrp.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbgrp_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bgrp.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbgrp_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bgrp.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_n_u8'}}
   return SVE_ACLE_FUNC(svbgrp,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bgrp.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbgrp_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bgrp.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbgrp_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bgrp.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_n_u16'}}
   return SVE_ACLE_FUNC(svbgrp,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bgrp.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbgrp_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bgrp.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbgrp_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bgrp.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_n_u32'}}
   return SVE_ACLE_FUNC(svbgrp,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svbgrp_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bgrp.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbgrp_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bgrp.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbgrp_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svbgrp_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bgrp.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbgrp'}}
   // expected-warning at +1 {{implicit declaration of function 'svbgrp_n_u64'}}
   return SVE_ACLE_FUNC(svbgrp,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl.c
index dc5fdd75f3bd7..dc75f24ecc99a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbsl_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svbsl_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svbsl_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_s8'}}
   return SVE_ACLE_FUNC(svbsl,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbsl_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svbsl_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_s16'}}
   return SVE_ACLE_FUNC(svbsl,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbsl_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svbsl_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_s32'}}
   return SVE_ACLE_FUNC(svbsl,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbsl_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svbsl_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_s64'}}
   return SVE_ACLE_FUNC(svbsl,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svbsl_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbsl_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_u8'}}
   return SVE_ACLE_FUNC(svbsl,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbsl_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbsl_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_u16'}}
   return SVE_ACLE_FUNC(svbsl,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbsl_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbsl_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_u32'}}
   return SVE_ACLE_FUNC(svbsl,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svbsl_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbsl_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_u64'}}
   return SVE_ACLE_FUNC(svbsl,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbsl_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbsl_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_s8'}}
   return SVE_ACLE_FUNC(svbsl,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svbsl_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_s16'}}
   return SVE_ACLE_FUNC(svbsl,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svbsl_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_s32'}}
   return SVE_ACLE_FUNC(svbsl,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svbsl_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_s64'}}
   return SVE_ACLE_FUNC(svbsl,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbsl_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbsl_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_u8'}}
   return SVE_ACLE_FUNC(svbsl,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbsl_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_u16'}}
   return SVE_ACLE_FUNC(svbsl,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbsl_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_u32'}}
   return SVE_ACLE_FUNC(svbsl,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbsl_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl_n_u64'}}
   return SVE_ACLE_FUNC(svbsl,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl1n.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl1n.c
index 15907841949b8..b3661c2c06900 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl1n.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl1n.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbsl1n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbsl1n_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svbsl1n_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_s8'}}
   return SVE_ACLE_FUNC(svbsl1n,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl1n_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svbsl1n_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_s16'}}
   return SVE_ACLE_FUNC(svbsl1n,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl1n_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svbsl1n_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_s32'}}
   return SVE_ACLE_FUNC(svbsl1n,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl1n_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svbsl1n_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_s64'}}
   return SVE_ACLE_FUNC(svbsl1n,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbsl1n_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbsl1n_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_u8'}}
   return SVE_ACLE_FUNC(svbsl1n,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl1n_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbsl1n_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_u16'}}
   return SVE_ACLE_FUNC(svbsl1n,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl1n_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbsl1n_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_u32'}}
   return SVE_ACLE_FUNC(svbsl1n,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl1n_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbsl1n_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_u64'}}
   return SVE_ACLE_FUNC(svbsl1n,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbsl1n_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbsl1n_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_s8'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl1n_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svbsl1n_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_s16'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl1n_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svbsl1n_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_s32'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl1n_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svbsl1n_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_s64'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbsl1n_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbsl1n_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_u8'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl1n_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbsl1n_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_u16'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl1n_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbsl1n_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_u32'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl1n_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl1n_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbsl1n_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl1n_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl1n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl1n_n_u64'}}
   return SVE_ACLE_FUNC(svbsl1n,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl2n.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl2n.c
index 67fb0d5381b51..1fc3dce061c7c 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl2n.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_bsl2n.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svbsl2n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbsl2n_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svbsl2n_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_s8'}}
   return SVE_ACLE_FUNC(svbsl2n,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl2n_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svbsl2n_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_s16'}}
   return SVE_ACLE_FUNC(svbsl2n,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl2n_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svbsl2n_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_s32'}}
   return SVE_ACLE_FUNC(svbsl2n,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl2n_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svbsl2n_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_s64'}}
   return SVE_ACLE_FUNC(svbsl2n,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svbsl2n_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svbsl2n_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_u8'}}
   return SVE_ACLE_FUNC(svbsl2n,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl2n_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svbsl2n_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_u16'}}
   return SVE_ACLE_FUNC(svbsl2n,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl2n_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svbsl2n_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_u32'}}
   return SVE_ACLE_FUNC(svbsl2n,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svbsl2n_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svbsl2n_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_u64'}}
   return SVE_ACLE_FUNC(svbsl2n,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbsl2n_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svbsl2n_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_s8'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl2n_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svbsl2n_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_s16'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl2n_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svbsl2n_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_s32'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl2n_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svbsl2n_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_s64'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svbsl2n_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svbsl2n_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_u8'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl2n_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svbsl2n_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_u16'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl2n_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svbsl2n_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_u32'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svbsl2n_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svbsl2n_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svbsl2n_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svbsl2n_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svbsl2n'}}
   // expected-warning at +1 {{implicit declaration of function 'svbsl2n_n_u64'}}
   return SVE_ACLE_FUNC(svbsl2n,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cadd.c
index 98a8b351ae21f..d506bb6946c20 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,161 +15,273 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcadd_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svcadd_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcadd_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 90)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s8'}}
   return SVE_ACLE_FUNC(svcadd,_s8,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcadd_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcadd_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 270)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s8'}}
   return SVE_ACLE_FUNC(svcadd,_s8,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcadd_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcadd_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s16'}}
   return SVE_ACLE_FUNC(svcadd,_s16,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcadd_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 270)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s16'}}
   return SVE_ACLE_FUNC(svcadd,_s16,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcadd_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcadd_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s32'}}
   return SVE_ACLE_FUNC(svcadd,_s32,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcadd_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s32'}}
   return SVE_ACLE_FUNC(svcadd,_s32,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcadd_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcadd_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s64'}}
   return SVE_ACLE_FUNC(svcadd,_s64,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcadd_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_s64'}}
   return SVE_ACLE_FUNC(svcadd,_s64,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svcadd_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcadd_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 90)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u8'}}
   return SVE_ACLE_FUNC(svcadd,_u8,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcadd_u8_1u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcadd_u8_1(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 270)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u8'}}
   return SVE_ACLE_FUNC(svcadd,_u8,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcadd_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcadd_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u16'}}
   return SVE_ACLE_FUNC(svcadd,_u16,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcadd_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 270)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u16'}}
   return SVE_ACLE_FUNC(svcadd,_u16,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcadd_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcadd_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u32'}}
   return SVE_ACLE_FUNC(svcadd,_u32,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcadd_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u32'}}
   return SVE_ACLE_FUNC(svcadd,_u32,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svcadd_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcadd_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svcadd_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u64'}}
   return SVE_ACLE_FUNC(svcadd,_u64,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svcadd_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcadd_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svcadd_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svcadd_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svcadd_u64'}}
   return SVE_ACLE_FUNC(svcadd,_u64,,)(op1, op2, 270);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cdot.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cdot.c
index ffc9f3be8d70e..fa23454b32998 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cdot.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cdot.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,111 +15,188 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcdot_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcdot_s32u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcdot_s32(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s32'}}
   return SVE_ACLE_FUNC(svcdot,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcdot_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcdot_s32_1u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcdot_s32_1(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s32'}}
   return SVE_ACLE_FUNC(svcdot,_s32,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcdot_s32_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcdot_s32_2u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcdot_s32_2(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s32_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 180)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s32'}}
   return SVE_ACLE_FUNC(svcdot,_s32,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcdot_s32_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcdot_s32_3u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcdot_s32_3(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s32_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s32'}}
   return SVE_ACLE_FUNC(svcdot,_s32,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcdot_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcdot_s64u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcdot_s64(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s64'}}
   return SVE_ACLE_FUNC(svcdot,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcdot_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcdot_s64_1u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcdot_s64_1(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s64'}}
   return SVE_ACLE_FUNC(svcdot,_s64,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcdot_s64_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcdot_s64_2u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcdot_s64_2(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s64_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 180)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s64'}}
   return SVE_ACLE_FUNC(svcdot,_s64,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcdot_s64_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcdot_s64_3u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcdot_s64_3(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcdot_s64_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_s64'}}
   return SVE_ACLE_FUNC(svcdot,_s64,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcdot_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0, i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcdot_lane_s32u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0, i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcdot_lane_s32(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcdot_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 0, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_lane_s32'}}
   return SVE_ACLE_FUNC(svcdot_lane,_s32,,)(op1, op2, op3, 0, 0);
 }
 
+// CHECK-LABEL: @test_svcdot_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 2, i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcdot_lane_s32_1u11__SVInt32_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 2, i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcdot_lane_s32_1(svint32_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcdot_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cdot.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 2, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_lane_s32'}}
   return SVE_ACLE_FUNC(svcdot_lane,_s32,,)(op1, op2, op3, 2, 90);
 }
 
+// CHECK-LABEL: @test_svcdot_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 180)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcdot_lane_s64u11__SVInt64_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcdot_lane_s64(svint64_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcdot_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cdot.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0, i32 180)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcdot_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svcdot_lane_s64'}}
   return SVE_ACLE_FUNC(svcdot_lane,_s64,,)(op1, op2, op3, 0, 180);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cmla.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cmla.c
index bf196a0b2dd6a..3dc907a46d656 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cmla.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cmla.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,362 +14,642 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcmla_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svcmla_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcmla_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s8'}}
   return SVE_ACLE_FUNC(svcmla,_s8,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmla_s8_1u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcmla_s8_1(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 90)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s8'}}
   return SVE_ACLE_FUNC(svcmla,_s8,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_s8_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmla_s8_2u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcmla_s8_2(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s8_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 180)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s8'}}
   return SVE_ACLE_FUNC(svcmla,_s8,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_s8_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmla_s8_3u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svcmla_s8_3(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s8_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 270)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s8'}}
   return SVE_ACLE_FUNC(svcmla,_s8,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmla_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcmla_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s16'}}
   return SVE_ACLE_FUNC(svcmla,_s16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcmla_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s16'}}
   return SVE_ACLE_FUNC(svcmla,_s16,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_s16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s16_2u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcmla_s16_2(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s16_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 180)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s16'}}
   return SVE_ACLE_FUNC(svcmla,_s16,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_s16_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s16_3u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcmla_s16_3(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s16_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 270)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s16'}}
   return SVE_ACLE_FUNC(svcmla,_s16,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmla_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcmla_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s32'}}
   return SVE_ACLE_FUNC(svcmla,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcmla_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s32'}}
   return SVE_ACLE_FUNC(svcmla,_s32,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_s32_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s32_2u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcmla_s32_2(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s32_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 180)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s32'}}
   return SVE_ACLE_FUNC(svcmla,_s32,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_s32_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s32_3u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcmla_s32_3(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s32_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s32'}}
   return SVE_ACLE_FUNC(svcmla,_s32,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmla_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcmla_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s64'}}
   return SVE_ACLE_FUNC(svcmla,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s64_1u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcmla_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s64'}}
   return SVE_ACLE_FUNC(svcmla,_s64,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_s64_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s64_2u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcmla_s64_2(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s64_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 180)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s64'}}
   return SVE_ACLE_FUNC(svcmla,_s64,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_s64_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_s64_3u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svcmla_s64_3(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_s64_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_s64'}}
   return SVE_ACLE_FUNC(svcmla,_s64,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svcmla_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcmla_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u8'}}
   return SVE_ACLE_FUNC(svcmla,_u8,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmla_u8_1u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcmla_u8_1(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 90)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u8'}}
   return SVE_ACLE_FUNC(svcmla,_u8,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_u8_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmla_u8_2u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcmla_u8_2(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u8_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 180)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u8'}}
   return SVE_ACLE_FUNC(svcmla,_u8,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_u8_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svcmla_u8_3u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svcmla_u8_3(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u8_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.cmla.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 270)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u8'}}
   return SVE_ACLE_FUNC(svcmla,_u8,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmla_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcmla_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u16'}}
   return SVE_ACLE_FUNC(svcmla,_u16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u16_1u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcmla_u16_1(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u16'}}
   return SVE_ACLE_FUNC(svcmla,_u16,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_u16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u16_2u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcmla_u16_2(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u16_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 180)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u16'}}
   return SVE_ACLE_FUNC(svcmla,_u16,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_u16_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u16_3u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcmla_u16_3(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u16_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 270)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u16'}}
   return SVE_ACLE_FUNC(svcmla,_u16,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmla_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcmla_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u32'}}
   return SVE_ACLE_FUNC(svcmla,_u32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u32_1u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcmla_u32_1(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u32'}}
   return SVE_ACLE_FUNC(svcmla,_u32,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_u32_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u32_2u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcmla_u32_2(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u32_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 180)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u32'}}
   return SVE_ACLE_FUNC(svcmla,_u32,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_u32_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u32_3u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcmla_u32_3(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u32_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u32'}}
   return SVE_ACLE_FUNC(svcmla,_u32,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svcmla_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svcmla_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u64'}}
   return SVE_ACLE_FUNC(svcmla,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u64_1u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svcmla_u64_1(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u64'}}
   return SVE_ACLE_FUNC(svcmla,_u64,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_u64_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u64_2u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svcmla_u64_2(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u64_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 180)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u64'}}
   return SVE_ACLE_FUNC(svcmla,_u64,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_u64_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svcmla_u64_3u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svcmla_u64_3(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svcmla_u64_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.cmla.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_u64'}}
   return SVE_ACLE_FUNC(svcmla,_u64,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmla_lane_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcmla_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_s16'}}
   return SVE_ACLE_FUNC(svcmla_lane,_s16,,)(op1, op2, op3, 0, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3, i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmla_lane_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3, i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svcmla_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 3, i32 180)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_s16'}}
   return SVE_ACLE_FUNC(svcmla_lane,_s16,,)(op1, op2, op3, 3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0, i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmla_lane_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0, i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcmla_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_s32'}}
   return SVE_ACLE_FUNC(svcmla_lane,_s32,,)(op1, op2, op3, 0, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 1, i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmla_lane_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 1, i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svcmla_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_s32'}}
   return SVE_ACLE_FUNC(svcmla_lane,_s32,,)(op1, op2, op3, 1, 0);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmla_lane_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcmla_lane_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_u16'}}
   return SVE_ACLE_FUNC(svcmla_lane,_u16,,)(op1, op2, op3, 0, 90);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3, i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmla_lane_u16_1u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3, i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svcmla_lane_u16_1(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.cmla.lane.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 3, i32 180)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_u16'}}
   return SVE_ACLE_FUNC(svcmla_lane,_u16,,)(op1, op2, op3, 3, 180);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0, i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcmla_lane_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0, i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcmla_lane_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_u32'}}
   return SVE_ACLE_FUNC(svcmla_lane,_u32,,)(op1, op2, op3, 0, 270);
 }
 
+// CHECK-LABEL: @test_svcmla_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 1, i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcmla_lane_u32_1u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 1, i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svcmla_lane_u32_1(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svcmla_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.cmla.lane.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svcmla_lane_u32'}}
   return SVE_ACLE_FUNC(svcmla_lane,_u32,,)(op1, op2, op3, 1, 0);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtlt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtlt.c
index e36fb406842a4..0a5c0ffb64eee 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtlt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtlt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,45 +15,77 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvtlt_f32_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtlt.f32f16(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtlt_f32_f16_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtlt.f32f16(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtlt_f32_f16_m(svfloat32_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvtlt_f32_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtlt.f32f16(<vscale x 4 x float> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtlt_f32_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtlt_f32_f16_m'}}
   return SVE_ACLE_FUNC(svcvtlt_f32,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtlt_f64_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvtlt.f64f32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtlt_f64_f32_mu13__SVFloat64_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvtlt.f64f32(<vscale x 2 x double> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvtlt_f64_f32_m(svfloat64_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvtlt_f64_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvtlt.f64f32(<vscale x 2 x double> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtlt_f64_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtlt_f64_f32_m'}}
   return SVE_ACLE_FUNC(svcvtlt_f64,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtlt_f32_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtlt.f32f16(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtlt_f32_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtlt.f32f16(<vscale x 4 x float> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtlt_f32_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svcvtlt_f32_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtlt.f32f16(<vscale x 4 x float> undef, <vscale x 4 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtlt_f32_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtlt_f32_f16_x'}}
   return SVE_ACLE_FUNC(svcvtlt_f32,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtlt_f64_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvtlt.f64f32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtlt_f64_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvtlt.f64f32(<vscale x 2 x double> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svcvtlt_f64_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvtlt_f64_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fcvtlt.f64f32(<vscale x 2 x double> undef, <vscale x 2 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtlt_f64_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtlt_f64_f32_x'}}
   return SVE_ACLE_FUNC(svcvtlt_f64,_f32,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtnt.c
index 2d45ad177a07e..d2080c6062c9b 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,45 +15,77 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvtnt_f16_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvtnt.f16f32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtnt_f16_f32_mu13__SVFloat16_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvtnt.f16f32(<vscale x 8 x half> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvtnt_f16_f32_m(svfloat16_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvtnt_f16_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvtnt.f16f32(<vscale x 8 x half> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtnt_f16_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtnt_f16_f32_m'}}
   return SVE_ACLE_FUNC(svcvtnt_f16,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtnt_f32_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtnt.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtnt_f32_f64_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtnt.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtnt_f32_f64_m(svfloat32_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtnt_f32_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtnt.f32f64(<vscale x 4 x float> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtnt_f32_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtnt_f32_f64_m'}}
   return SVE_ACLE_FUNC(svcvtnt_f32,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtnt_f16_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvtnt.f16f32(<vscale x 8 x half> [[EVEN:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtnt_f16_f32_xu13__SVFloat16_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvtnt.f16f32(<vscale x 8 x half> [[EVEN:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svcvtnt_f16_f32_x(svfloat16_t even, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svcvtnt_f16_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fcvtnt.f16f32(<vscale x 8 x half> %even, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtnt_f16_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtnt_f16_f32_x'}}
   return SVE_ACLE_FUNC(svcvtnt_f16,_f32,_x,)(even, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtnt_f32_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtnt.f32f64(<vscale x 4 x float> [[EVEN:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svcvtnt_f32_f64_xu13__SVFloat32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtnt.f32f64(<vscale x 4 x float> [[EVEN:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtnt_f32_f64_x(svfloat32_t even, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtnt_f32_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtnt.f32f64(<vscale x 4 x float> %even, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtnt_f32_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtnt_f32_f64_x'}}
   return SVE_ACLE_FUNC(svcvtnt_f32,_f64,_x,)(even, pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtx.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtx.c
index a482db147c23a..65ed353f42257 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtx.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtx.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,34 +15,58 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvtx_f32_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcvtx_f32_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtx_f32_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtx_f32_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtx_f32_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtx_f32_f64_z'}}
   return SVE_ACLE_FUNC(svcvtx_f32,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtx_f32_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcvtx_f32_f64_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtx_f32_f64_m(svfloat32_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtx_f32_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtx_f32_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtx_f32_f64_m'}}
   return SVE_ACLE_FUNC(svcvtx_f32,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtx_f32_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svcvtx_f32_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtx_f32_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtx_f32_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtx.f32f64(<vscale x 4 x float> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtx_f32_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtx_f32_f64_x'}}
   return SVE_ACLE_FUNC(svcvtx_f32,_f64,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtxnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtxnt.c
index aeb2a7b758830..acac4ab436814 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtxnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_cvtxnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,23 +15,39 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svcvtxnt_f32_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtxnt.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcvtxnt_f32_f64_mu13__SVFloat32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtxnt.f32f64(<vscale x 4 x float> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtxnt_f32_f64_m(svfloat32_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtxnt_f32_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtxnt.f32f64(<vscale x 4 x float> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtxnt_f32_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtxnt_f32_f64_m'}}
   return SVE_ACLE_FUNC(svcvtxnt_f32,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svcvtxnt_f32_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtxnt.f32f64(<vscale x 4 x float> [[EVEN:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svcvtxnt_f32_f64_xu13__SVFloat32_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtxnt.f32f64(<vscale x 4 x float> [[EVEN:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svcvtxnt_f32_f64_x(svfloat32_t even, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svcvtxnt_f32_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fcvtxnt.f32f64(<vscale x 4 x float> %even, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svcvtxnt_f32_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svcvtxnt_f32_f64_x'}}
   return SVE_ACLE_FUNC(svcvtxnt_f32,_f64,_x,)(even, pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eor3.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eor3.c
index 5efa4fd866628..bcb020225ad2e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eor3.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eor3.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_sveor3_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_sveor3_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_sveor3_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_sveor3_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_s8'}}
   return SVE_ACLE_FUNC(sveor3,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor3_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_sveor3_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_sveor3_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_s16'}}
   return SVE_ACLE_FUNC(sveor3,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor3_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_sveor3_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_sveor3_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_s32'}}
   return SVE_ACLE_FUNC(sveor3,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor3_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_sveor3_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_sveor3_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_s64'}}
   return SVE_ACLE_FUNC(sveor3,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_sveor3_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_sveor3_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_sveor3_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_u8'}}
   return SVE_ACLE_FUNC(sveor3,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor3_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_sveor3_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_sveor3_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_u16'}}
   return SVE_ACLE_FUNC(sveor3,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor3_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_sveor3_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_sveor3_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_u32'}}
   return SVE_ACLE_FUNC(sveor3,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveor3_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_sveor3_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_sveor3_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_u64'}}
   return SVE_ACLE_FUNC(sveor3,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor3_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_sveor3_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_s8'}}
   return SVE_ACLE_FUNC(sveor3,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor3_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_sveor3_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_s16'}}
   return SVE_ACLE_FUNC(sveor3,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor3_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_sveor3_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_s32'}}
   return SVE_ACLE_FUNC(sveor3,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor3_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_sveor3_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_s64'}}
   return SVE_ACLE_FUNC(sveor3,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveor3_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_sveor3_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_u8'}}
   return SVE_ACLE_FUNC(sveor3,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor3_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_sveor3_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_u16'}}
   return SVE_ACLE_FUNC(sveor3,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor3_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_sveor3_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_u32'}}
   return SVE_ACLE_FUNC(sveor3,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveor3_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveor3_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_sveor3_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_sveor3_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveor3'}}
   // expected-warning at +1 {{implicit declaration of function 'sveor3_n_u64'}}
   return SVE_ACLE_FUNC(sveor3,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eorbt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eorbt.c
index 262d41fbd355a..c813ebc049ea1 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eorbt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eorbt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_sveorbt_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorbt_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_sveorbt_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_s8'}}
   return SVE_ACLE_FUNC(sveorbt,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveorbt_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_sveorbt_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_s16'}}
   return SVE_ACLE_FUNC(sveorbt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveorbt_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_sveorbt_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_s32'}}
   return SVE_ACLE_FUNC(sveorbt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveorbt_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_sveorbt_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_s64'}}
   return SVE_ACLE_FUNC(sveorbt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveorbt_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_sveorbt_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_u8'}}
   return SVE_ACLE_FUNC(sveorbt,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveorbt_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_sveorbt_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_u16'}}
   return SVE_ACLE_FUNC(sveorbt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveorbt_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_sveorbt_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_u32'}}
   return SVE_ACLE_FUNC(sveorbt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveorbt_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_sveorbt_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_u64'}}
   return SVE_ACLE_FUNC(sveorbt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveorbt_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_sveorbt_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_s8'}}
   return SVE_ACLE_FUNC(sveorbt,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveorbt_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_sveorbt_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_s16'}}
   return SVE_ACLE_FUNC(sveorbt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveorbt_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_sveorbt_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_s32'}}
   return SVE_ACLE_FUNC(sveorbt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveorbt_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_sveorbt_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_s64'}}
   return SVE_ACLE_FUNC(sveorbt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveorbt_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_sveorbt_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eorbt.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_u8'}}
   return SVE_ACLE_FUNC(sveorbt,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveorbt_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_sveorbt_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eorbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_u16'}}
   return SVE_ACLE_FUNC(sveorbt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveorbt_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_sveorbt_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eorbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_u32'}}
   return SVE_ACLE_FUNC(sveorbt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveorbt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveorbt_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_sveorbt_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_sveorbt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eorbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveorbt'}}
   // expected-warning at +1 {{implicit declaration of function 'sveorbt_n_u64'}}
   return SVE_ACLE_FUNC(sveorbt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eortb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eortb.c
index 0b305f15ce768..fdcac8f2edd25 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eortb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_eortb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_sveortb_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveortb_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_sveortb_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_sveortb_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_s8'}}
   return SVE_ACLE_FUNC(sveortb,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveortb_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_sveortb_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_sveortb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_s16'}}
   return SVE_ACLE_FUNC(sveortb,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveortb_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_sveortb_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_sveortb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_s32'}}
   return SVE_ACLE_FUNC(sveortb,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveortb_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_sveortb_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_sveortb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_s64'}}
   return SVE_ACLE_FUNC(sveortb,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_sveortb_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_sveortb_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_sveortb_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_u8'}}
   return SVE_ACLE_FUNC(sveortb,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveortb_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_sveortb_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_sveortb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_u16'}}
   return SVE_ACLE_FUNC(sveortb,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveortb_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_sveortb_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_sveortb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_u32'}}
   return SVE_ACLE_FUNC(sveortb,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_sveortb_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_sveortb_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_sveortb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_u64'}}
   return SVE_ACLE_FUNC(sveortb,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveortb_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_sveortb_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_s8'}}
   return SVE_ACLE_FUNC(sveortb,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveortb_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_sveortb_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_s16'}}
   return SVE_ACLE_FUNC(sveortb,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveortb_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_sveortb_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_s32'}}
   return SVE_ACLE_FUNC(sveortb,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveortb_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_sveortb_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_s64'}}
   return SVE_ACLE_FUNC(sveortb,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_sveortb_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_sveortb_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.eortb.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_u8'}}
   return SVE_ACLE_FUNC(sveortb,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveortb_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_sveortb_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.eortb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_u16'}}
   return SVE_ACLE_FUNC(sveortb,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveortb_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_sveortb_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.eortb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_u32'}}
   return SVE_ACLE_FUNC(sveortb,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_sveortb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_sveortb_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_sveortb_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_sveortb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.eortb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'sveortb'}}
   // expected-warning at +1 {{implicit declaration of function 'sveortb_n_u64'}}
   return SVE_ACLE_FUNC(sveortb,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hadd.c
index 9f70269dfea38..bdcc297bbe840 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,972 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svhadd_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhadd_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svhadd_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s8_m'}}
   return SVE_ACLE_FUNC(svhadd,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svhadd_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s16_m'}}
   return SVE_ACLE_FUNC(svhadd,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svhadd_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s32_m'}}
   return SVE_ACLE_FUNC(svhadd,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svhadd_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s64_m'}}
   return SVE_ACLE_FUNC(svhadd,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhadd_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhadd_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u8_m'}}
   return SVE_ACLE_FUNC(svhadd,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svhadd_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u16_m'}}
   return SVE_ACLE_FUNC(svhadd,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhadd_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svhadd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u32_m'}}
   return SVE_ACLE_FUNC(svhadd,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhadd_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u64_m'}}
   return SVE_ACLE_FUNC(svhadd,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhadd_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhadd_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s8_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhadd_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s16_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhadd_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s32_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhadd_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s64_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhadd_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhadd_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u8_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhadd_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u16_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhadd_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u32_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhadd_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u64_m'}}
   return SVE_ACLE_FUNC(svhadd,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhadd_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhadd_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s8_z'}}
   return SVE_ACLE_FUNC(svhadd,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhadd_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s16_z'}}
   return SVE_ACLE_FUNC(svhadd,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhadd_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s32_z'}}
   return SVE_ACLE_FUNC(svhadd,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhadd_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s64_z'}}
   return SVE_ACLE_FUNC(svhadd,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhadd_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhadd_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u8_z'}}
   return SVE_ACLE_FUNC(svhadd,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhadd_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u16_z'}}
   return SVE_ACLE_FUNC(svhadd,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhadd_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svhadd_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u32_z'}}
   return SVE_ACLE_FUNC(svhadd,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhadd_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u64_z'}}
   return SVE_ACLE_FUNC(svhadd,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhadd_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svhadd_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s8_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svhadd_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s16_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svhadd_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s32_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svhadd_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s64_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhadd_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svhadd_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u8_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svhadd_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u16_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svhadd_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u32_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svhadd_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u64_z'}}
   return SVE_ACLE_FUNC(svhadd,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhadd_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svhadd_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s8_x'}}
   return SVE_ACLE_FUNC(svhadd,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svhadd_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s16_x'}}
   return SVE_ACLE_FUNC(svhadd,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svhadd_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s32_x'}}
   return SVE_ACLE_FUNC(svhadd,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svhadd_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_s64_x'}}
   return SVE_ACLE_FUNC(svhadd,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhadd_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhadd_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u8_x'}}
   return SVE_ACLE_FUNC(svhadd,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svhadd_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u16_x'}}
   return SVE_ACLE_FUNC(svhadd,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhadd_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svhadd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u32_x'}}
   return SVE_ACLE_FUNC(svhadd,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhadd_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhadd_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_u64_x'}}
   return SVE_ACLE_FUNC(svhadd,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhadd_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhadd_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s8_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhadd_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s16_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhadd_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s32_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhadd_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_s64_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhadd_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhadd_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u8_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhadd_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u16_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhadd_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u32_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhadd_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhadd_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhadd_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhadd_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhadd_n_u64_x'}}
   return SVE_ACLE_FUNC(svhadd,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histcnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histcnt.c
index 1161b6a91836d..9e8a98f3fd145 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histcnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histcnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,45 +15,77 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svhistcnt_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhistcnt_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhistcnt_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhistcnt_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhistcnt_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhistcnt_s32_z'}}
   return SVE_ACLE_FUNC(svhistcnt,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhistcnt_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhistcnt_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhistcnt_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhistcnt_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhistcnt_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhistcnt_s64_z'}}
   return SVE_ACLE_FUNC(svhistcnt,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhistcnt_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhistcnt_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhistcnt_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhistcnt_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.histcnt.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhistcnt_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhistcnt_u32_z'}}
   return SVE_ACLE_FUNC(svhistcnt,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhistcnt_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhistcnt_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhistcnt_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhistcnt_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.histcnt.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhistcnt_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhistcnt_u64_z'}}
   return SVE_ACLE_FUNC(svhistcnt,_u64,_z,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histseg.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histseg.c
index e49a3993197d6..acb672eb773f8 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histseg.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_histseg.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,21 +15,35 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svhistseg_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.histseg.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhistseg_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.histseg.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhistseg_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhistseg_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.histseg.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhistseg'}}
   // expected-warning at +1 {{implicit declaration of function 'svhistseg_s8'}}
   return SVE_ACLE_FUNC(svhistseg,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svhistseg_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.histseg.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhistseg_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.histseg.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhistseg_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhistseg_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.histseg.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhistseg'}}
   // expected-warning at +1 {{implicit declaration of function 'svhistseg_u8'}}
   return SVE_ACLE_FUNC(svhistseg,_u8,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsub.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsub.c
index efce49325e2b3..47bdeea578370 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsub.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsub.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svhsub_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhsub_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhsub_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s8_z'}}
   return SVE_ACLE_FUNC(svhsub,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhsub_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s16_z'}}
   return SVE_ACLE_FUNC(svhsub,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhsub_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s32_z'}}
   return SVE_ACLE_FUNC(svhsub,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhsub_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s64_z'}}
   return SVE_ACLE_FUNC(svhsub,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhsub_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhsub_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u8_z'}}
   return SVE_ACLE_FUNC(svhsub,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhsub_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u16_z'}}
   return SVE_ACLE_FUNC(svhsub,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhsub_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u32_z'}}
   return SVE_ACLE_FUNC(svhsub,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhsub_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u64_z'}}
   return SVE_ACLE_FUNC(svhsub,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhsub_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svhsub_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s8_m'}}
   return SVE_ACLE_FUNC(svhsub,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svhsub_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s16_m'}}
   return SVE_ACLE_FUNC(svhsub,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svhsub_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s32_m'}}
   return SVE_ACLE_FUNC(svhsub,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svhsub_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s64_m'}}
   return SVE_ACLE_FUNC(svhsub,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhsub_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhsub_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u8_m'}}
   return SVE_ACLE_FUNC(svhsub,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svhsub_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u16_m'}}
   return SVE_ACLE_FUNC(svhsub,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhsub_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u32_m'}}
   return SVE_ACLE_FUNC(svhsub,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhsub_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u64_m'}}
   return SVE_ACLE_FUNC(svhsub,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhsub_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svhsub_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s8_x'}}
   return SVE_ACLE_FUNC(svhsub,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svhsub_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s16_x'}}
   return SVE_ACLE_FUNC(svhsub,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svhsub_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s32_x'}}
   return SVE_ACLE_FUNC(svhsub,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svhsub_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_s64_x'}}
   return SVE_ACLE_FUNC(svhsub,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svhsub_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhsub_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u8_x'}}
   return SVE_ACLE_FUNC(svhsub,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svhsub_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u16_x'}}
   return SVE_ACLE_FUNC(svhsub,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhsub_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u32_x'}}
   return SVE_ACLE_FUNC(svhsub,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsub_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhsub_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_u64_x'}}
   return SVE_ACLE_FUNC(svhsub,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsub_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svhsub_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s8_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svhsub_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s16_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svhsub_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s32_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svhsub_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s64_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsub_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svhsub_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u8_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svhsub_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u16_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svhsub_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u32_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svhsub_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u64_z'}}
   return SVE_ACLE_FUNC(svhsub,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsub_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhsub_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s8_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhsub_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s16_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhsub_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s32_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhsub_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s64_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsub_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhsub_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u8_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhsub_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u16_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhsub_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u32_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhsub_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u64_m'}}
   return SVE_ACLE_FUNC(svhsub,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsub_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhsub_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s8_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhsub_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s16_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhsub_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s32_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhsub_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_s64_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsub_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhsub_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u8_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhsub_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u16_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhsub_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u32_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsub_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsub_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhsub_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhsub_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsub_n_u64_x'}}
   return SVE_ACLE_FUNC(svhsub,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsubr.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsubr.c
index 381b760db8544..823cfd8681757 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsubr.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_hsubr.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,556 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svhsubr_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsubr_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhsubr_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s8_z'}}
   return SVE_ACLE_FUNC(svhsubr,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhsubr_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s16_z'}}
   return SVE_ACLE_FUNC(svhsubr,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhsubr_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s32_z'}}
   return SVE_ACLE_FUNC(svhsubr,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhsubr_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s64_z'}}
   return SVE_ACLE_FUNC(svhsubr,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsubr_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhsubr_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u8_z'}}
   return SVE_ACLE_FUNC(svhsubr,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhsubr_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u16_z'}}
   return SVE_ACLE_FUNC(svhsubr,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhsubr_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u32_z'}}
   return SVE_ACLE_FUNC(svhsubr,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhsubr_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u64_z'}}
   return SVE_ACLE_FUNC(svhsubr,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsubr_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svhsubr_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s8_m'}}
   return SVE_ACLE_FUNC(svhsubr,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svhsubr_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s16_m'}}
   return SVE_ACLE_FUNC(svhsubr,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svhsubr_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s32_m'}}
   return SVE_ACLE_FUNC(svhsubr,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svhsubr_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s64_m'}}
   return SVE_ACLE_FUNC(svhsubr,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsubr_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhsubr_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u8_m'}}
   return SVE_ACLE_FUNC(svhsubr,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svhsubr_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u16_m'}}
   return SVE_ACLE_FUNC(svhsubr,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhsubr_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u32_m'}}
   return SVE_ACLE_FUNC(svhsubr,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhsubr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u64_m'}}
   return SVE_ACLE_FUNC(svhsubr,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsubr_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svhsubr_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s8_x'}}
   return SVE_ACLE_FUNC(svhsubr,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svhsubr_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s16_x'}}
   return SVE_ACLE_FUNC(svhsubr,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svhsubr_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s32_x'}}
   return SVE_ACLE_FUNC(svhsubr,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svhsubr_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_s64_x'}}
   return SVE_ACLE_FUNC(svhsubr,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svhsubr_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svhsubr_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u8_x'}}
   return SVE_ACLE_FUNC(svhsubr,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svhsubr_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u16_x'}}
   return SVE_ACLE_FUNC(svhsubr,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svhsubr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u32_x'}}
   return SVE_ACLE_FUNC(svhsubr,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svhsubr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svhsubr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_u64_x'}}
   return SVE_ACLE_FUNC(svhsubr,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsubr_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svhsubr_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s8_z
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s8_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svhsubr_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s16_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svhsubr_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s32_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svhsubr_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s64_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsubr_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svhsubr_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u8_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svhsubr_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u16_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svhsubr_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u32_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svhsubr_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u64_z'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsubr_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhsubr_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s8_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhsubr_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s16_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhsubr_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s32_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhsubr_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s64_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsubr_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhsubr_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u8_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhsubr_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u16_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhsubr_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u32_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhsubr_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u64_m'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsubr_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svhsubr_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s8_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svhsubr_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s16_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svhsubr_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s32_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svhsubr_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.shsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_s64_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svhsubr_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svhsubr_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uhsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u8_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svhsubr_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uhsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u16_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svhsubr_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uhsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u32_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svhsubr_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svhsubr_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svhsubr_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svhsubr_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uhsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svhsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svhsubr_n_u64_x'}}
   return SVE_ACLE_FUNC(svhsubr,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1.c
index 47febd9823a4b..5c9b61009a49a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,337 +15,607 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldnt1_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldnt1_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_s32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldnt1_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldnt1_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldnt1_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_u32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldnt1_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldnt1_gather_u32base_f32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldnt1_gather_u32base_f32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_f32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_f32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _f32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z31test_svldnt1_gather_u64base_f64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_gather_u64base_f64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_f64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _f64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_s64offset_s64u10__SVBool_tPKlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_gather_s64offset_s64(svbool_t pg, const int64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, s64, offset, _s64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_s64offset_u64u10__SVBool_tPKmu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_gather_s64offset_u64(svbool_t pg, const uint64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, s64, offset, _u64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_s64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_s64offset_f64u10__SVBool_tPKdu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_gather_s64offset_f64(svbool_t pg, const float64_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_s64offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_s64offset_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, s64, offset, _f64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_u32offset_s32u10__SVBool_tPKiu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldnt1_gather_u32offset_s32(svbool_t pg, const int32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u32, offset, _s32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_u64offset_s64u10__SVBool_tPKlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_gather_u64offset_s64(svbool_t pg, const int64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u64, offset, _s64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_u32offset_u32u10__SVBool_tPKju12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldnt1_gather_u32offset_u32(svbool_t pg, const uint32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u32, offset, _u32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_u64offset_u64u10__SVBool_tPKmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_gather_u64offset_u64(svbool_t pg, const uint64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u64, offset, _u64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_u32offset_f32u10__SVBool_tPKfu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldnt1_gather_u32offset_f32(svbool_t pg, const float32_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_u32offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1> [[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32offset_f32'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u32, offset, _f32)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1_gather_u64offset_f64u10__SVBool_tPKdu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_gather_u64offset_f64(svbool_t pg, const float64_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1_gather_u64offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64offset_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u64, offset, _f64)(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldnt1_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svldnt1_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldnt1_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldnt1_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svldnt1_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldnt1_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldnt1_gather_u32base_offset_f32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svldnt1_gather_u32base_offset_f32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset_f32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_offset_f32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _offset_f32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z38test_svldnt1_gather_u64base_offset_f64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_gather_u64base_offset_f64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_offset_f64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_offset_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _offset_f64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldnt1_gather_s64index_s64u10__SVBool_tPKlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_gather_s64index_s64(svbool_t pg, const int64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_s64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, s64, index, _s64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldnt1_gather_s64index_u64u10__SVBool_tPKmu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_gather_s64index_u64(svbool_t pg, const uint64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_s64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, s64, index, _u64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_s64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldnt1_gather_s64index_f64u10__SVBool_tPKdu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_gather_s64index_f64(svbool_t pg, const float64_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1_gather_s64index_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.index.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_s64index_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, s64, index, _f64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldnt1_gather_u64index_s64u10__SVBool_tPKlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svldnt1_gather_u64index_s64(svbool_t pg, const int64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u64, index, _s64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldnt1_gather_u64index_u64u10__SVBool_tPKmu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svldnt1_gather_u64index_u64(svbool_t pg, const uint64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i64(<vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u64, index, _u64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z32test_svldnt1_gather_u64index_f64u10__SVBool_tPKdu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.index.nxv2f64(<vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svldnt1_gather_u64index_f64(svbool_t pg, const float64_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1_gather_u64index_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.index.nxv2f64(<vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64index_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather_, u64, index, _f64)(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldnt1_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_index_s32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldnt1_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldnt1_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 4 x i32> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_index_u32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldnt1_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 2 x i64> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _index_u64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u32base_index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldnt1_gather_u32base_index_f32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svldnt1_gather_u32base_index_f32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1_gather_u32base_index_f32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 4 x float> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index_f32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u32base_index_f32'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u32base, _index_f32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1_gather_u64base_index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z37test_svldnt1_gather_u64base_index_f64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svldnt1_gather_u64base_index_f64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1_gather_u64base_index_f64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret <vscale x 2 x double> [[LOAD]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1_gather_index_f64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1_gather_u64base_index_f64'}}
   return SVE_ACLE_FUNC(svldnt1_gather, _u64base, _index_f64, )(pg, bases, index);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sb.c
index 44164b1b21a79..a031b754fe7ef 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,155 +15,281 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sb_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1sb_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u32base_s32'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sb_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sb_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sb_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1sb_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u32base_u32'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sb_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sb_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sb_gather_s64offset_s64u10__SVBool_tPKau11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sb_gather_s64offset_s64(svbool_t pg, const int8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sb_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sb_gather_s64offset_u64u10__SVBool_tPKau11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sb_gather_s64offset_u64(svbool_t pg, const int8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sb_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sb_gather_u32offset_s32u10__SVBool_tPKau12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1sb_gather_u32offset_s32(svbool_t pg, const int8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u32offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sb_gather_u64offset_s64u10__SVBool_tPKau12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sb_gather_u64offset_s64(svbool_t pg, const int8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sb_gather_u32offset_u32u10__SVBool_tPKau12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1sb_gather_u32offset_u32(svbool_t pg, const int8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u32offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sb_gather_u64offset_u64u10__SVBool_tPKau12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sb_gather_u64offset_u64(svbool_t pg, const int8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sb_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1sb_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sb_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sb_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sb_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1sb_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sb_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sb_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sb_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sb_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sb_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sb_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sb_gather, _u64base, _offset_u64, )(pg, bases, offset);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sh.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sh.c
index ad439a224fce9..09a700210e2c8 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sh.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sh.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,247 +15,449 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sh_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1sh_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32base_s32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sh_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sh_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sh_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1sh_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32base_u32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sh_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sh_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sh_gather_s64offset_s64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sh_gather_s64offset_s64(svbool_t pg, const int16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sh_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sh_gather_s64offset_u64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sh_gather_s64offset_u64(svbool_t pg, const int16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sh_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sh_gather_u32offset_s32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1sh_gather_u32offset_s32(svbool_t pg, const int16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sh_gather_u64offset_s64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sh_gather_u64offset_s64(svbool_t pg, const int16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sh_gather_u32offset_u32u10__SVBool_tPKsu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1sh_gather_u32offset_u32(svbool_t pg, const int16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sh_gather_u64offset_u64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sh_gather_u64offset_u64(svbool_t pg, const int16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sh_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1sh_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sh_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sh_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sh_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1sh_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sh_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sh_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sh_gather_s64index_s64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sh_gather_s64index_s64(svbool_t pg, const int16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sh_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_s64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sh_gather_s64index_u64u10__SVBool_tPKsu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sh_gather_s64index_u64(svbool_t pg, const int16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sh_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_s64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sh_gather_u64index_s64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sh_gather_u64index_s64(svbool_t pg, const int16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sh_gather_u64index_u64u10__SVBool_tPKsu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sh_gather_u64index_u64(svbool_t pg, const int16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1sh_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldnt1sh_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32base_index_s32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1sh_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldnt1sh_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1sh_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldnt1sh_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u32base_index_u32'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1sh_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1sh_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldnt1sh_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1sh_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sh_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sh_gather_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svldnt1sh_gather, _u64base, _index_u64, )(pg, bases, index);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sw.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sw.c
index 96dcbd2d94209..ec5ca9de59cf7 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sw.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1sw.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,157 +15,285 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sw_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sw_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1sw_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sw_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sw_gather_s64offset_s64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sw_gather_s64offset_s64(svbool_t pg, const int32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sw_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sw_gather_s64offset_u64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sw_gather_s64offset_u64(svbool_t pg, const int32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sw_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sw_gather_u64offset_s64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sw_gather_u64offset_s64(svbool_t pg, const int32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1sw_gather_u64offset_u64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sw_gather_u64offset_u64(svbool_t pg, const int32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sw_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sw_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1sw_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sw_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sw_gather_s64index_s64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sw_gather_s64index_s64(svbool_t pg, const int32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sw_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_s64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sw_gather_s64index_u64u10__SVBool_tPKiu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sw_gather_s64index_u64(svbool_t pg, const int32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sw_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_s64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sw_gather_u64index_s64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1sw_gather_u64index_s64(svbool_t pg, const int32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1sw_gather_u64index_u64u10__SVBool_tPKiu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = sext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1sw_gather_u64index_u64(svbool_t pg, const int32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1sw_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldnt1sw_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1sw_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1sw_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = sext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldnt1sw_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1sw_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[SEXT:%.*]] = sext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[SEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1sw_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1sw_gather_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svldnt1sw_gather, _u64base, _index_u64, )(pg, bases, index);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1ub.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1ub.c
index 29454ffb7ecca..e7e2fd8a63f88 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1ub.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1ub.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,155 +15,281 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1ub_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1ub_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u32base_s32'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1ub_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1ub_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1ub_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1ub_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u32base_u32'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1ub_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1ub_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1ub_gather_s64offset_s64u10__SVBool_tPKhu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1ub_gather_s64offset_s64(svbool_t pg, const uint8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1ub_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1ub_gather_s64offset_u64u10__SVBool_tPKhu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1ub_gather_s64offset_u64(svbool_t pg, const uint8_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1ub_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1ub_gather_u32offset_s32u10__SVBool_tPKhu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1ub_gather_u32offset_s32(svbool_t pg, const uint8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u32offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1ub_gather_u64offset_s64u10__SVBool_tPKhu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1ub_gather_u64offset_s64(svbool_t pg, const uint8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1ub_gather_u32offset_u32u10__SVBool_tPKhu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1ub_gather_u32offset_u32(svbool_t pg, const uint8_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u32offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1ub_gather_u64offset_u64u10__SVBool_tPKhu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[TMP0]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1ub_gather_u64offset_u64(svbool_t pg, const uint8_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.nxv2i8(<vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1ub_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1ub_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1ub_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1ub_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1ub_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i8> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1ub_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i8> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1ub_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1ub_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i8> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1ub_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1ub_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i8> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i8> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1ub_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1ub_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1ub_gather, _u64base, _offset_u64, )(pg, bases, offset);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uh.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uh.c
index f9b9161df295d..5080daa286dd0 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uh.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uh.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,247 +15,449 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1uh_gather_u32base_s32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1uh_gather_u32base_s32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32base_s32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u32base, _s32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1uh_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uh_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1uh_gather_u32base_u32u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1uh_gather_u32base_u32(svbool_t pg, svuint32_t bases) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32base_u32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u32base, _u32, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1uh_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uh_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uh_gather_s64offset_s64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uh_gather_s64offset_s64(svbool_t pg, const uint16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uh_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uh_gather_s64offset_u64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uh_gather_s64offset_u64(svbool_t pg, const uint16_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uh_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uh_gather_u32offset_s32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1uh_gather_u32offset_s32(svbool_t pg, const uint16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, u32, offset_s32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uh_gather_u64offset_s64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uh_gather_u64offset_s64(svbool_t pg, const uint16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uh_gather_u32offset_u32u10__SVBool_tPKtu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1uh_gather_u32offset_u32(svbool_t pg, const uint16_t *base, svuint32_t offsets) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, u32, offset_u32, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uh_gather_u64offset_u64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uh_gather_u64offset_u64(svbool_t pg, const uint16_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1uh_gather_u32base_offset_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svldnt1uh_gather_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u32base, _offset_s32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1uh_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uh_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1uh_gather_u32base_offset_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 4 x i16> [[TMP1]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svldnt1uh_gather_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u32base, _offset_u32, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1uh_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uh_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uh_gather_s64index_s64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uh_gather_s64index_s64(svbool_t pg, const uint16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uh_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_s64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uh_gather_s64index_u64u10__SVBool_tPKtu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uh_gather_s64index_u64(svbool_t pg, const uint16_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uh_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_s64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uh_gather_u64index_s64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uh_gather_u64index_s64(svbool_t pg, const uint16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uh_gather_u64index_u64u10__SVBool_tPKtu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[TMP0]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i16> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uh_gather_u64index_u64(svbool_t pg, const uint16_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i16(<vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1uh_gather_u32base_index_s32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svldnt1uh_gather_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_s32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32base_index_s32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u32base, _index_s32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1uh_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldnt1uh_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1uh_gather_u32base_index_u32u10__SVBool_tu12__SVUint32_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 4 x i16> [[TMP2]] to <vscale x 4 x i32>
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svldnt1uh_gather_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 4 x i16> [[LOAD]] to <vscale x 4 x i32>
-  // CHECK: ret <vscale x 4 x i32> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_u32'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u32base_index_u32'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u32base, _index_u32, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1uh_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1uh_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i16> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldnt1uh_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1uh_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i16> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i16> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uh_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uh_gather_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svldnt1uh_gather, _u64base, _index_u64, )(pg, bases, index);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uw.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uw.c
index cf95768e0a717..2480c09e770be 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uw.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_ldnt1uw.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,157 +15,285 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1uw_gather_u64base_s64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uw_gather_u64base_s64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64base_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather, _u64base, _s64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z33test_svldnt1uw_gather_u64base_u64u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uw_gather_u64base_u64(svbool_t pg, svuint64_t bases) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64base_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather, _u64base, _u64, )(pg, bases);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uw_gather_s64offset_s64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uw_gather_s64offset_s64(svbool_t pg, const uint32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uw_gather_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_s64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, s64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uw_gather_s64offset_u64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uw_gather_s64offset_u64(svbool_t pg, const uint32_t *base, svint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uw_gather_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_s64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, s64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uw_gather_u64offset_s64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uw_gather_u64offset_s64(svbool_t pg, const uint32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, u64, offset_s64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z35test_svldnt1uw_gather_u64offset_u64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uw_gather_u64offset_u64(svbool_t pg, const uint32_t *base, svuint64_t offsets) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, u64, offset_u64, )(pg, base, offsets);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1uw_gather_u64base_offset_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uw_gather_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_offset_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather, _u64base, _offset_s64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z40test_svldnt1uw_gather_u64base_offset_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uw_gather_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_offset_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather, _u64base, _offset_u64, )(pg, bases, offset);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uw_gather_s64index_s64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uw_gather_s64index_s64(svbool_t pg, const uint32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uw_gather_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_s64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, s64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uw_gather_s64index_u64u10__SVBool_tPKju11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uw_gather_s64index_u64(svbool_t pg, const uint32_t *base, svint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uw_gather_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_s64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, s64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uw_gather_u64index_s64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svldnt1uw_gather_u64index_s64(svbool_t pg, const uint32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64index_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, u64, index_s64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z34test_svldnt1uw_gather_u64index_u64u10__SVBool_tPKju12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = zext <vscale x 2 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svldnt1uw_gather_u64index_u64(svbool_t pg, const uint32_t *base, svuint64_t indices) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.index.nxv2i32(<vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64index_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather_, u64, index_u64, )(pg, base, indices);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1uw_gather_u64base_index_s64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svldnt1uw_gather_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_index_s64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather, _u64base, _index_s64, )(pg, bases, index);
 }
 
+// CHECK-LABEL: @test_svldnt1uw_gather_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z39test_svldnt1uw_gather_u64base_index_u64u10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = zext <vscale x 2 x i32> [[TMP2]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svldnt1uw_gather_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index) {
-  // CHECK-LABEL: test_svldnt1uw_gather_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: [[LOAD:%.*]] = call <vscale x 2 x i32> @llvm.aarch64.sve.ldnt1.gather.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: [[ZEXT:%.*]] = zext <vscale x 2 x i32> [[LOAD]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> [[ZEXT]]
   // overload-warning at +2 {{implicit declaration of function 'svldnt1uw_gather_index_u64'}}
   // expected-warning at +1 {{implicit declaration of function 'svldnt1uw_gather_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svldnt1uw_gather, _u64base, _index_u64, )(pg, bases, index);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_logb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_logb.c
index bcb227120ed46..35e7707a64cbe 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_logb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_logb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,100 +15,172 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svlogb_f16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f16_zu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlogb_f16_z(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svlogb_f16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f16_z'}}
   return SVE_ACLE_FUNC(svlogb,_f16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f32_zu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlogb_f32_z(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svlogb_f32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f32_z'}}
   return SVE_ACLE_FUNC(svlogb,_f32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f64_zu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svlogb_f64_z(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svlogb_f64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f64_z'}}
   return SVE_ACLE_FUNC(svlogb,_f64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f16_mu11__SVInt16_tu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlogb_f16_m(svint16_t inactive, svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svlogb_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f16_m'}}
   return SVE_ACLE_FUNC(svlogb,_f16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f32_mu11__SVInt32_tu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlogb_f32_m(svint32_t inactive, svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svlogb_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f32_m'}}
   return SVE_ACLE_FUNC(svlogb,_f32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f64_mu11__SVInt64_tu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svlogb_f64_m(svint64_t inactive, svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svlogb_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f64_m'}}
   return SVE_ACLE_FUNC(svlogb,_f64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f16_xu10__SVBool_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svlogb_f16_x(svbool_t pg, svfloat16_t op)
 {
-  // CHECK-LABEL: test_svlogb_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.flogb.nxv8f16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f16_x'}}
   return SVE_ACLE_FUNC(svlogb,_f16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f32_xu10__SVBool_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svlogb_f32_x(svbool_t pg, svfloat32_t op)
 {
-  // CHECK-LABEL: test_svlogb_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.flogb.nxv4f32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f32_x'}}
   return SVE_ACLE_FUNC(svlogb,_f32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svlogb_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svlogb_f64_xu10__SVBool_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svlogb_f64_x(svbool_t pg, svfloat64_t op)
 {
-  // CHECK-LABEL: test_svlogb_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.flogb.nxv2f64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svlogb_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svlogb_f64_x'}}
   return SVE_ACLE_FUNC(svlogb,_f64,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_match.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_match.c
index d98bd678a9023..2346b9dfcf9ca 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_match.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_match.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,45 +15,77 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmatch_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.match.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmatch_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.match.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svmatch_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmatch_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.match.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svmatch_s8'}}
   return SVE_ACLE_FUNC(svmatch,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmatch_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.match.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmatch_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.match.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svmatch_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmatch_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.match.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[RET:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[RET]]
   // overload-warning at +2 {{implicit declaration of function 'svmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svmatch_s16'}}
   return SVE_ACLE_FUNC(svmatch,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmatch_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.match.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svmatch_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.match.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svmatch_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmatch_u8
-  // CHECK: %[[intrinsic:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.match.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[intrinsic]]
   // overload-warning at +2 {{implicit declaration of function 'svmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svmatch_u8'}}
   return SVE_ACLE_FUNC(svmatch,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmatch_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.match.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmatch_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.match.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svmatch_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmatch_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.match.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[RET:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[RET]]
   // overload-warning at +2 {{implicit declaration of function 'svmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svmatch_u16'}}
   return SVE_ACLE_FUNC(svmatch,_u16,,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxnmp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxnmp.c
index 01e43f95fdd50..c32a22afbf3de 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxnmp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxnmp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,67 +15,115 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmaxnmp_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmaxnmp_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmaxnmp_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnmp_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnmp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxnmp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxnmp_f16_m'}}
   return SVE_ACLE_FUNC(svmaxnmp,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnmp_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmaxnmp_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmaxnmp_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnmp_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnmp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxnmp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxnmp_f32_m'}}
   return SVE_ACLE_FUNC(svmaxnmp,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnmp_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmaxnmp_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmaxnmp_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnmp_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnmp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxnmp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxnmp_f64_m'}}
   return SVE_ACLE_FUNC(svmaxnmp,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnmp_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmaxnmp_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmaxnmp_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxnmp_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxnmp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxnmp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxnmp_f16_x'}}
   return SVE_ACLE_FUNC(svmaxnmp,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnmp_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmaxnmp_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmaxnmp_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxnmp_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxnmp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxnmp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxnmp_f32_x'}}
   return SVE_ACLE_FUNC(svmaxnmp,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxnmp_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmaxnmp_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmaxnmp_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxnmp_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxnmp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxnmp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxnmp_f64_x'}}
   return SVE_ACLE_FUNC(svmaxnmp,_f64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxp.c
index fcf84455cb391..8339579289096 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_maxp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,239 +15,411 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmaxp_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmaxp_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmaxp_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smaxp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s8_m'}}
   return SVE_ACLE_FUNC(svmaxp,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmaxp_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smaxp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s16_m'}}
   return SVE_ACLE_FUNC(svmaxp,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmaxp_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smaxp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s32_m'}}
   return SVE_ACLE_FUNC(svmaxp,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmaxp_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smaxp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s64_m'}}
   return SVE_ACLE_FUNC(svmaxp,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmaxp_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmaxp_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umaxp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u8_m'}}
   return SVE_ACLE_FUNC(svmaxp,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmaxp_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umaxp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u16_m'}}
   return SVE_ACLE_FUNC(svmaxp,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmaxp_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umaxp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u32_m'}}
   return SVE_ACLE_FUNC(svmaxp,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmaxp_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umaxp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u64_m'}}
   return SVE_ACLE_FUNC(svmaxp,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmaxp_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svmaxp_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.smaxp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s8_x'}}
   return SVE_ACLE_FUNC(svmaxp,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmaxp_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smaxp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s16_x'}}
   return SVE_ACLE_FUNC(svmaxp,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmaxp_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smaxp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s32_x'}}
   return SVE_ACLE_FUNC(svmaxp,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmaxp_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smaxp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_s64_x'}}
   return SVE_ACLE_FUNC(svmaxp,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmaxp_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umaxp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svmaxp_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.umaxp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u8_x'}}
   return SVE_ACLE_FUNC(svmaxp,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umaxp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmaxp_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umaxp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u16_x'}}
   return SVE_ACLE_FUNC(svmaxp,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umaxp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmaxp_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umaxp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u32_x'}}
   return SVE_ACLE_FUNC(svmaxp,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umaxp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmaxp_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umaxp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_u64_x'}}
   return SVE_ACLE_FUNC(svmaxp,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmaxp_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_f16_m'}}
   return SVE_ACLE_FUNC(svmaxp,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmaxp_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_f32_m'}}
   return SVE_ACLE_FUNC(svmaxp,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmaxp_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_f64_m'}}
   return SVE_ACLE_FUNC(svmaxp,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svmaxp_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fmaxp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_f16_x'}}
   return SVE_ACLE_FUNC(svmaxp,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmaxp_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmaxp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_f32_x'}}
   return SVE_ACLE_FUNC(svmaxp,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svmaxp_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svmaxp_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svmaxp_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svmaxp_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fmaxp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmaxp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svmaxp_f64_x'}}
   return SVE_ACLE_FUNC(svmaxp,_f64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minnmp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minnmp.c
index 4072511e3d524..cef87cdf88462 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minnmp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minnmp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,67 +15,115 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svminnmp_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svminnmp_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svminnmp_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminnmp_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnmp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminnmp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminnmp_f16_m'}}
   return SVE_ACLE_FUNC(svminnmp,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnmp_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svminnmp_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svminnmp_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminnmp_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnmp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminnmp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminnmp_f32_m'}}
   return SVE_ACLE_FUNC(svminnmp,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnmp_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svminnmp_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svminnmp_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminnmp_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnmp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminnmp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminnmp_f64_m'}}
   return SVE_ACLE_FUNC(svminnmp,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnmp_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svminnmp_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnmp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svminnmp_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminnmp_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminnmp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminnmp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminnmp_f16_x'}}
   return SVE_ACLE_FUNC(svminnmp,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnmp_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svminnmp_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnmp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svminnmp_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminnmp_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminnmp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminnmp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminnmp_f32_x'}}
   return SVE_ACLE_FUNC(svminnmp,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminnmp_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svminnmp_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnmp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svminnmp_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminnmp_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminnmp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminnmp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminnmp_f64_x'}}
   return SVE_ACLE_FUNC(svminnmp,_f64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minp.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minp.c
index 9885b2b936f64..ac0c62f94cf68 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minp.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_minp.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,239 +15,411 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svminp_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svminp_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svminp_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svminp_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sminp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s8_m'}}
   return SVE_ACLE_FUNC(svminp,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svminp_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svminp_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sminp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s16_m'}}
   return SVE_ACLE_FUNC(svminp,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svminp_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svminp_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sminp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s32_m'}}
   return SVE_ACLE_FUNC(svminp,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svminp_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svminp_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sminp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s64_m'}}
   return SVE_ACLE_FUNC(svminp,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svminp_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svminp_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svminp_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uminp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u8_m'}}
   return SVE_ACLE_FUNC(svminp,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svminp_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svminp_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uminp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u16_m'}}
   return SVE_ACLE_FUNC(svminp,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svminp_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svminp_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uminp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u32_m'}}
   return SVE_ACLE_FUNC(svminp,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svminp_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svminp_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uminp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u64_m'}}
   return SVE_ACLE_FUNC(svminp,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svminp_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svminp_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svminp_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sminp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s8_x'}}
   return SVE_ACLE_FUNC(svminp,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svminp_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svminp_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sminp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s16_x'}}
   return SVE_ACLE_FUNC(svminp,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svminp_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svminp_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sminp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s32_x'}}
   return SVE_ACLE_FUNC(svminp,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svminp_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svminp_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sminp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_s64_x'}}
   return SVE_ACLE_FUNC(svminp,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svminp_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uminp.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svminp_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svminp_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uminp.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u8_x'}}
   return SVE_ACLE_FUNC(svminp,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uminp.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svminp_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svminp_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uminp.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u16_x'}}
   return SVE_ACLE_FUNC(svminp,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uminp.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svminp_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svminp_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uminp.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u32_x'}}
   return SVE_ACLE_FUNC(svminp,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uminp.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svminp_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svminp_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uminp.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_u64_x'}}
   return SVE_ACLE_FUNC(svminp,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_f16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_f16_mu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svminp_f16_m(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminp_f16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_f16_m'}}
   return SVE_ACLE_FUNC(svminp,_f16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_f32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_f32_mu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svminp_f32_m(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminp_f32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_f32_m'}}
   return SVE_ACLE_FUNC(svminp,_f32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_f64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_f64_mu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svminp_f64_m(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminp_f64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_f64_m'}}
   return SVE_ACLE_FUNC(svminp,_f64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_f16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_f16_xu10__SVBool_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminp.nxv8f16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x half> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP1]]
+//
 svfloat16_t test_svminp_f16_x(svbool_t pg, svfloat16_t op1, svfloat16_t op2)
 {
-  // CHECK-LABEL: test_svminp_f16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.fminp.nxv8f16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x half> %op1, <vscale x 8 x half> %op2)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_f16_x'}}
   return SVE_ACLE_FUNC(svminp,_f16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_f32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_f32_xu10__SVBool_tu13__SVFloat32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminp.nxv4f32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svminp_f32_x(svbool_t pg, svfloat32_t op1, svfloat32_t op2)
 {
-  // CHECK-LABEL: test_svminp_f32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fminp.nxv4f32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x float> %op1, <vscale x 4 x float> %op2)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_f32_x'}}
   return SVE_ACLE_FUNC(svminp,_f32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svminp_f64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svminp_f64_xu10__SVBool_tu13__SVFloat64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminp.nxv2f64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP1]]
+//
 svfloat64_t test_svminp_f64_x(svbool_t pg, svfloat64_t op1, svfloat64_t op2)
 {
-  // CHECK-LABEL: test_svminp_f64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.fminp.nxv2f64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x double> %op1, <vscale x 2 x double> %op2)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svminp_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svminp_f64_x'}}
   return SVE_ACLE_FUNC(svminp,_f64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mla.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mla.c
index c6e866ab9e145..1b6cb83fa3d73 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mla.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mla.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,101 +14,178 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmla_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmla_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_s16'}}
   return SVE_ACLE_FUNC(svmla_lane,_s16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmla_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_s16'}}
   return SVE_ACLE_FUNC(svmla_lane,_s16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmla_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmla_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_s32'}}
   return SVE_ACLE_FUNC(svmla_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmla_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_s32'}}
   return SVE_ACLE_FUNC(svmla_lane,_s32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmla_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmla_lane_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_s64'}}
   return SVE_ACLE_FUNC(svmla_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_s64_1u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmla_lane_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_s64'}}
   return SVE_ACLE_FUNC(svmla_lane,_s64,,)(op1, op2, op3, 1);
 }
 
+// CHECK-LABEL: @test_svmla_lane_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmla_lane_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_u16'}}
   return SVE_ACLE_FUNC(svmla_lane,_u16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_u16_1u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmla_lane_u16_1(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mla.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_u16'}}
   return SVE_ACLE_FUNC(svmla_lane,_u16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmla_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_u32_1u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmla_lane_u32_1(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mla.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_u32'}}
   return SVE_ACLE_FUNC(svmla_lane,_u32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmla_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmla_lane_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmla_lane_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_u64'}}
   return SVE_ACLE_FUNC(svmla_lane,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmla_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmla_lane_u64_1u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmla_lane_u64_1(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmla_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mla.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmla_lane_u64'}}
   return SVE_ACLE_FUNC(svmla_lane,_u64,,)(op1, op2, op3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalb.c
index 9efa5f168f558..f3385a5408167 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,248 +15,423 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmlalb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmlalb_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_s16'}}
   return SVE_ACLE_FUNC(svmlalb,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlalb_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_s32'}}
   return SVE_ACLE_FUNC(svmlalb,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlalb_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_s64'}}
   return SVE_ACLE_FUNC(svmlalb,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_u16u12__SVUint16_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmlalb_u16(svuint16_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_u16'}}
   return SVE_ACLE_FUNC(svmlalb,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlalb_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_u32'}}
   return SVE_ACLE_FUNC(svmlalb,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlalb_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_u64'}}
   return SVE_ACLE_FUNC(svmlalb,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmlalb_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_s16'}}
   return SVE_ACLE_FUNC(svmlalb,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmlalb_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_s32'}}
   return SVE_ACLE_FUNC(svmlalb,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmlalb_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_s64'}}
   return SVE_ACLE_FUNC(svmlalb,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_u16u12__SVUint16_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmlalb_n_u16(svuint16_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_u16'}}
   return SVE_ACLE_FUNC(svmlalb,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_u32u12__SVUint32_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmlalb_n_u32(svuint32_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_u32'}}
   return SVE_ACLE_FUNC(svmlalb,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_u64u12__SVUint64_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmlalb_n_u64(svuint64_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_u64'}}
   return SVE_ACLE_FUNC(svmlalb,_n_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalb_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlalb_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_s32'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalb_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlalb_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_s32'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalb_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlalb_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_s64'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalb_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlalb_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_s64'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_s64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalb_lane_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlalb_lane_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_u32'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_u32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalb_lane_u32_1u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlalb_lane_u32_1(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_u32'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_u32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalb_lane_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlalb_lane_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_u64'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalb_lane_u64_1u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlalb_lane_u64_1(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_u64'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_u64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlalb_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalb_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlalb_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_f32'}}
   return SVE_ACLE_FUNC(svmlalb,_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalb_n_f32u13__SVFloat32_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmlalb_n_f32(svfloat32_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_n_f32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_n_f32'}}
   return SVE_ACLE_FUNC(svmlalb,_n_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalb_lane_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlalb_lane_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_f32'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_f32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalb_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalb_lane_f32_1u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlalb_lane_f32_1(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlalb_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalb.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalb_lane_f32'}}
   return SVE_ACLE_FUNC(svmlalb_lane,_f32,,)(op1, op2, op3, 7);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalt.c
index 95d3eac914c82..a5c75f811a7d5 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlalt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,248 +15,423 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmlalt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmlalt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_s16'}}
   return SVE_ACLE_FUNC(svmlalt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlalt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_s32'}}
   return SVE_ACLE_FUNC(svmlalt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlalt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_s64'}}
   return SVE_ACLE_FUNC(svmlalt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_u16u12__SVUint16_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmlalt_u16(svuint16_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_u16'}}
   return SVE_ACLE_FUNC(svmlalt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlalt_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_u32'}}
   return SVE_ACLE_FUNC(svmlalt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlalt_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_u64'}}
   return SVE_ACLE_FUNC(svmlalt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmlalt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_s16'}}
   return SVE_ACLE_FUNC(svmlalt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmlalt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_s32'}}
   return SVE_ACLE_FUNC(svmlalt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmlalt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_s64'}}
   return SVE_ACLE_FUNC(svmlalt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_u16u12__SVUint16_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmlalt_n_u16(svuint16_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_u16'}}
   return SVE_ACLE_FUNC(svmlalt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_u32u12__SVUint32_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmlalt_n_u32(svuint32_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_u32'}}
   return SVE_ACLE_FUNC(svmlalt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_u64u12__SVUint64_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmlalt_n_u64(svuint64_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_u64'}}
   return SVE_ACLE_FUNC(svmlalt,_n_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalt_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlalt_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_s32'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalt_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlalt_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlalt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_s32'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalt_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlalt_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_s64'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalt_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlalt_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlalt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_s64'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_s64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalt_lane_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlalt_lane_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_u32'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_u32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalt_lane_u32_1u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlalt_lane_u32_1(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlalt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_u32'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_u32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalt_lane_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlalt_lane_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_u64'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalt_lane_u64_1u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlalt_lane_u64_1(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlalt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_u64'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_u64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlalt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlalt_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlalt_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_f32'}}
   return SVE_ACLE_FUNC(svmlalt,_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlalt_n_f32u13__SVFloat32_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmlalt_n_f32(svfloat32_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_n_f32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_n_f32'}}
   return SVE_ACLE_FUNC(svmlalt,_n_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlalt_lane_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlalt_lane_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_f32'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_f32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlalt_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlalt_lane_f32_1u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlalt_lane_f32_1(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlalt_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlalt.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlalt_lane_f32'}}
   return SVE_ACLE_FUNC(svmlalt_lane,_f32,,)(op1, op2, op3, 7);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mls.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mls.c
index c9370f1580014..60fec481b5aa6 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mls.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mls.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,110 +14,194 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmls_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmls_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_s16'}}
   return SVE_ACLE_FUNC(svmls_lane,_s16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmls_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_s16'}}
   return SVE_ACLE_FUNC(svmls_lane,_s16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmls_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmls_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_s32'}}
   return SVE_ACLE_FUNC(svmls_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmls_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_s32'}}
   return SVE_ACLE_FUNC(svmls_lane,_s32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmls_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmls_lane_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_s64'}}
   return SVE_ACLE_FUNC(svmls_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_s64_1u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmls_lane_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_s64'}}
   return SVE_ACLE_FUNC(svmls_lane,_s64,,)(op1, op2, op3, 1);
 }
 
+// CHECK-LABEL: @test_svmls_lane_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmls_lane_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_u16'}}
   return SVE_ACLE_FUNC(svmls_lane,_u16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_u16_1u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmls_lane_u16_1(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mls.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_u16'}}
   return SVE_ACLE_FUNC(svmls_lane,_u16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmls_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmls_lane_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_u32'}}
   return SVE_ACLE_FUNC(svmls_lane,_u32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_u32_1u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmls_lane_u32_1(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mls.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_u32'}}
   return SVE_ACLE_FUNC(svmls_lane,_u32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmls_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmls_lane_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmls_lane_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_u64'}}
   return SVE_ACLE_FUNC(svmls_lane,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmls_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmls_lane_u64_1u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmls_lane_u64_1(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svmls_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mls.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmls_lane_u64'}}
   return SVE_ACLE_FUNC(svmls_lane,_u64,,)(op1, op2, op3, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslb.c
index f96e1d9dba4c7..457abe252c3f5 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,248 +15,423 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmlslb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmlslb_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_s16'}}
   return SVE_ACLE_FUNC(svmlslb,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlslb_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_s32'}}
   return SVE_ACLE_FUNC(svmlslb,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlslb_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_s64'}}
   return SVE_ACLE_FUNC(svmlslb,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_u16u12__SVUint16_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmlslb_u16(svuint16_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_u16'}}
   return SVE_ACLE_FUNC(svmlslb,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlslb_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_u32'}}
   return SVE_ACLE_FUNC(svmlslb,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlslb_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_u64'}}
   return SVE_ACLE_FUNC(svmlslb,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmlslb_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_s16'}}
   return SVE_ACLE_FUNC(svmlslb,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmlslb_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_s32'}}
   return SVE_ACLE_FUNC(svmlslb,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmlslb_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_s64'}}
   return SVE_ACLE_FUNC(svmlslb,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_u16u12__SVUint16_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmlslb_n_u16(svuint16_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_u16'}}
   return SVE_ACLE_FUNC(svmlslb,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_u32u12__SVUint32_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmlslb_n_u32(svuint32_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_u32'}}
   return SVE_ACLE_FUNC(svmlslb,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_u64u12__SVUint64_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmlslb_n_u64(svuint64_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_u64'}}
   return SVE_ACLE_FUNC(svmlslb,_n_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslb_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlslb_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_s32'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslb_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlslb_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_s32'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslb_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlslb_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_s64'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslb_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlslb_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_s64'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_s64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslb_lane_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlslb_lane_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_u32'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_u32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslb_lane_u32_1u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlslb_lane_u32_1(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_u32'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_u32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslb_lane_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlslb_lane_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_u64'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslb_lane_u64_1u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlslb_lane_u64_1(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_u64'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_u64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlslb_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslb_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlslb_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_f32'}}
   return SVE_ACLE_FUNC(svmlslb,_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslb_n_f32u13__SVFloat32_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmlslb_n_f32(svfloat32_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_n_f32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_n_f32'}}
   return SVE_ACLE_FUNC(svmlslb,_n_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslb_lane_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlslb_lane_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_f32'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_f32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslb_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslb_lane_f32_1u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlslb_lane_f32_1(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlslb_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslb.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslb_lane_f32'}}
   return SVE_ACLE_FUNC(svmlslb_lane,_f32,,)(op1, op2, op3, 7);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslt.c
index 458ddd26c2a37..66ee032b91b13 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mlslt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,248 +15,423 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmlslt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmlslt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_s16'}}
   return SVE_ACLE_FUNC(svmlslt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlslt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_s32'}}
   return SVE_ACLE_FUNC(svmlslt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlslt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_s64'}}
   return SVE_ACLE_FUNC(svmlslt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_u16u12__SVUint16_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmlslt_u16(svuint16_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_u16'}}
   return SVE_ACLE_FUNC(svmlslt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlslt_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_u32'}}
   return SVE_ACLE_FUNC(svmlslt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlslt_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_u64'}}
   return SVE_ACLE_FUNC(svmlslt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmlslt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smlslt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_s16'}}
   return SVE_ACLE_FUNC(svmlslt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmlslt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_s32'}}
   return SVE_ACLE_FUNC(svmlslt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmlslt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_s64'}}
   return SVE_ACLE_FUNC(svmlslt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_u16u12__SVUint16_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmlslt_n_u16(svuint16_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umlslt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_u16'}}
   return SVE_ACLE_FUNC(svmlslt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_u32u12__SVUint32_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmlslt_n_u32(svuint32_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_u32'}}
   return SVE_ACLE_FUNC(svmlslt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_u64u12__SVUint64_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmlslt_n_u64(svuint64_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_u64'}}
   return SVE_ACLE_FUNC(svmlslt,_n_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslt_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlslt_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_s32'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslt_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmlslt_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smlslt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_s32'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslt_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlslt_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_s64'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslt_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmlslt_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smlslt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_s64'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_s64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslt_lane_u32u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlslt_lane_u32(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_u32'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_u32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslt_lane_u32_1u12__SVUint32_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmlslt_lane_u32_1(svuint32_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umlslt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_u32'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_u32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslt_lane_u64u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlslt_lane_u64(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_u64'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_u64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslt_lane_u64_1u12__SVUint64_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmlslt_lane_u64_1(svuint64_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umlslt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_u64'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_u64,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svmlslt_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmlslt_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlslt_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_f32'}}
   return SVE_ACLE_FUNC(svmlslt,_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_n_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmlslt_n_f32u13__SVFloat32_tu13__SVFloat16_tDh(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP1]]
+//
 svfloat32_t test_svmlslt_n_f32(svfloat32_t op1, svfloat16_t op2, float16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_n_f32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.dup.x.nxv8f16(half %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %[[DUP]])
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_n_f32'}}
   return SVE_ACLE_FUNC(svmlslt,_n_f32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmlslt_lane_f32u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlslt_lane_f32(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_f32'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_f32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svmlslt_lane_f32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmlslt_lane_f32_1u13__SVFloat32_tu13__SVFloat16_tu13__SVFloat16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.lane.nxv4f32(<vscale x 4 x float> [[OP1:%.*]], <vscale x 8 x half> [[OP2:%.*]], <vscale x 8 x half> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svmlslt_lane_f32_1(svfloat32_t op1, svfloat16_t op2, svfloat16_t op3)
 {
-  // CHECK-LABEL: test_svmlslt_lane_f32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.fmlslt.lane.nxv4f32(<vscale x 4 x float> %op1, <vscale x 8 x half> %op2, <vscale x 8 x half> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmlslt_lane_f32'}}
   return SVE_ACLE_FUNC(svmlslt_lane,_f32,,)(op1, op2, op3, 7);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlb.c
index 0fbffffcc771b..2157d147e5156 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmovlb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlb_s16u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmovlb_s16(svint8_t op1)
 {
-  // CHECK-LABEL: test_svmovlb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlb_s16'}}
   return SVE_ACLE_FUNC(svmovlb,_s16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlb_s32u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmovlb_s32(svint16_t op1)
 {
-  // CHECK-LABEL: test_svmovlb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlb_s32'}}
   return SVE_ACLE_FUNC(svmovlb,_s32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlb_s64u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmovlb_s64(svint32_t op1)
 {
-  // CHECK-LABEL: test_svmovlb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlb_s64'}}
   return SVE_ACLE_FUNC(svmovlb,_s64,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlb_u16u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmovlb_u16(svuint8_t op1)
 {
-  // CHECK-LABEL: test_svmovlb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlb_u16'}}
   return SVE_ACLE_FUNC(svmovlb,_u16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlb_u32u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmovlb_u32(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svmovlb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlb_u32'}}
   return SVE_ACLE_FUNC(svmovlb,_u32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlb_u64u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmovlb_u64(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svmovlb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlb_u64'}}
   return SVE_ACLE_FUNC(svmovlb,_u64,,)(op1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlt.c
index 74c2187f6fccf..c7ceb8655fc7d 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_movlt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmovlt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlt_s16u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmovlt_s16(svint8_t op1)
 {
-  // CHECK-LABEL: test_svmovlt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlt_s16'}}
   return SVE_ACLE_FUNC(svmovlt,_s16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlt_s32u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmovlt_s32(svint16_t op1)
 {
-  // CHECK-LABEL: test_svmovlt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlt_s32'}}
   return SVE_ACLE_FUNC(svmovlt,_s32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlt_s64u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmovlt_s64(svint32_t op1)
 {
-  // CHECK-LABEL: test_svmovlt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlt_s64'}}
   return SVE_ACLE_FUNC(svmovlt,_s64,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlt_u16u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmovlt_u16(svuint8_t op1)
 {
-  // CHECK-LABEL: test_svmovlt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlt_u16'}}
   return SVE_ACLE_FUNC(svmovlt,_u16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlt_u32u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmovlt_u32(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svmovlt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlt_u32'}}
   return SVE_ACLE_FUNC(svmovlt,_u32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svmovlt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmovlt_u64u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmovlt_u64(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svmovlt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmovlt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmovlt_u64'}}
   return SVE_ACLE_FUNC(svmovlt,_u64,,)(op1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mul.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mul.c
index bd4559e1e9eaf..eeda5b494b602 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mul.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mul.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
 
 #include <arm_sve.h>
@@ -13,110 +14,194 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmul_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmul_lane_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_s16'}}
   return SVE_ACLE_FUNC(svmul_lane,_s16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmul_lane_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_s16'}}
   return SVE_ACLE_FUNC(svmul_lane,_s16,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmul_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmul_lane_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_s32'}}
   return SVE_ACLE_FUNC(svmul_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmul_lane_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_s32'}}
   return SVE_ACLE_FUNC(svmul_lane,_s32,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svmul_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmul_lane_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_s64'}}
   return SVE_ACLE_FUNC(svmul_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmul_lane_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_s64'}}
   return SVE_ACLE_FUNC(svmul_lane,_s64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svmul_lane_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmul_lane_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_u16'}}
   return SVE_ACLE_FUNC(svmul_lane,_u16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmul_lane_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.mul.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_u16'}}
   return SVE_ACLE_FUNC(svmul_lane,_u16,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmul_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmul_lane_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_u32'}}
   return SVE_ACLE_FUNC(svmul_lane,_u32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmul_lane_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.mul.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_u32'}}
   return SVE_ACLE_FUNC(svmul_lane,_u32,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svmul_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svmul_lane_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmul_lane_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_u64'}}
   return SVE_ACLE_FUNC(svmul_lane,_u64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmul_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmul_lane_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmul_lane_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svmul_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.mul.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // expected-warning at +1 {{implicit declaration of function 'svmul_lane_u64'}}
   return SVE_ACLE_FUNC(svmul_lane,_u64,,)(op1, op2, 1);
 }

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullb.c
index c8f6166e708d4..70723deb57bf0 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,207 +15,353 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmullb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullb_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmullb_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmullb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_s16'}}
   return SVE_ACLE_FUNC(svmullb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullb_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmullb_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_s32'}}
   return SVE_ACLE_FUNC(svmullb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullb_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmullb_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_s64'}}
   return SVE_ACLE_FUNC(svmullb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullb_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmullb_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmullb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_u16'}}
   return SVE_ACLE_FUNC(svmullb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullb_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmullb_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_u32'}}
   return SVE_ACLE_FUNC(svmullb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullb_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmullb_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_u64'}}
   return SVE_ACLE_FUNC(svmullb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullb_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmullb_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmullb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_n_s16'}}
   return SVE_ACLE_FUNC(svmullb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullb_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmullb_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_n_s32'}}
   return SVE_ACLE_FUNC(svmullb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullb_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmullb_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_n_s64'}}
   return SVE_ACLE_FUNC(svmullb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullb_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmullb_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmullb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_n_u16'}}
   return SVE_ACLE_FUNC(svmullb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullb_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmullb_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_n_u32'}}
   return SVE_ACLE_FUNC(svmullb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullb_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmullb_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_n_u64'}}
   return SVE_ACLE_FUNC(svmullb,_n_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullb_lane_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmullb_lane_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_s32'}}
   return SVE_ACLE_FUNC(svmullb_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullb_lane_s32_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmullb_lane_s32_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullb.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_s32'}}
   return SVE_ACLE_FUNC(svmullb_lane,_s32,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullb_lane_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmullb_lane_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_s64'}}
   return SVE_ACLE_FUNC(svmullb_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullb_lane_s64_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmullb_lane_s64_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullb.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_s64'}}
   return SVE_ACLE_FUNC(svmullb_lane,_s64,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullb_lane_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmullb_lane_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_u32'}}
   return SVE_ACLE_FUNC(svmullb_lane,_u32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullb_lane_u32_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmullb_lane_u32_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullb.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_u32'}}
   return SVE_ACLE_FUNC(svmullb_lane,_u32,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullb_lane_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmullb_lane_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_u64'}}
   return SVE_ACLE_FUNC(svmullb_lane,_u64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullb_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullb_lane_u64_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmullb_lane_u64_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmullb_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullb.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullb_lane_u64'}}
   return SVE_ACLE_FUNC(svmullb_lane,_u64,,)(op1, op2, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullt.c
index 13ac0f34e905c..b02989e681e5b 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_mullt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,207 +15,353 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svmullt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svmullt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svmullt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_s16'}}
   return SVE_ACLE_FUNC(svmullt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmullt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_s32'}}
   return SVE_ACLE_FUNC(svmullt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmullt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_s64'}}
   return SVE_ACLE_FUNC(svmullt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullt_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svmullt_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svmullt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_u16'}}
   return SVE_ACLE_FUNC(svmullt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullt_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmullt_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_u32'}}
   return SVE_ACLE_FUNC(svmullt,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svmullt_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmullt_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_u64'}}
   return SVE_ACLE_FUNC(svmullt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svmullt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svmullt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.smullt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_n_s16'}}
   return SVE_ACLE_FUNC(svmullt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svmullt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_n_s32'}}
   return SVE_ACLE_FUNC(svmullt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svmullt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_n_s64'}}
   return SVE_ACLE_FUNC(svmullt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullt_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svmullt_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svmullt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.umullt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_n_u16'}}
   return SVE_ACLE_FUNC(svmullt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullt_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svmullt_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_n_u32'}}
   return SVE_ACLE_FUNC(svmullt,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svmullt_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svmullt_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_n_u64'}}
   return SVE_ACLE_FUNC(svmullt,_n_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullt_lane_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmullt_lane_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_s32'}}
   return SVE_ACLE_FUNC(svmullt_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullt_lane_s32_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svmullt_lane_s32_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.smullt.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_s32'}}
   return SVE_ACLE_FUNC(svmullt_lane,_s32,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullt_lane_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmullt_lane_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_s64'}}
   return SVE_ACLE_FUNC(svmullt_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullt_lane_s64_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svmullt_lane_s64_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.smullt.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_s64'}}
   return SVE_ACLE_FUNC(svmullt_lane,_s64,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullt_lane_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmullt_lane_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_u32'}}
   return SVE_ACLE_FUNC(svmullt_lane,_u32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullt_lane_u32_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svmullt_lane_u32_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.umullt.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_u32'}}
   return SVE_ACLE_FUNC(svmullt_lane,_u32,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svmullt_lane_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmullt_lane_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_u64'}}
   return SVE_ACLE_FUNC(svmullt_lane,_u64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svmullt_lane_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svmullt_lane_u64_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svmullt_lane_u64_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svmullt_lane_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.umullt.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svmullt_lane_u64'}}
   return SVE_ACLE_FUNC(svmullt_lane,_u64,,)(op1, op2, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nbsl.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nbsl.c
index 444606b295a03..54e5a9314a2a6 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nbsl.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nbsl.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,169 +15,289 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnbsl_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svnbsl_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svnbsl_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_s8'}}
   return SVE_ACLE_FUNC(svnbsl,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnbsl_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svnbsl_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_s16'}}
   return SVE_ACLE_FUNC(svnbsl,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnbsl_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svnbsl_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_s32'}}
   return SVE_ACLE_FUNC(svnbsl,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnbsl_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svnbsl_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_s64'}}
   return SVE_ACLE_FUNC(svnbsl,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svnbsl_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svnbsl_u8(svuint8_t op1, svuint8_t op2, svuint8_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_u8'}}
   return SVE_ACLE_FUNC(svnbsl,_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnbsl_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svnbsl_u16(svuint16_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_u16'}}
   return SVE_ACLE_FUNC(svnbsl,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnbsl_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svnbsl_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_u32'}}
   return SVE_ACLE_FUNC(svnbsl,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svnbsl_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svnbsl_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_u64'}}
   return SVE_ACLE_FUNC(svnbsl,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnbsl_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svnbsl_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_s8'}}
   return SVE_ACLE_FUNC(svnbsl,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnbsl_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svnbsl_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_s16'}}
   return SVE_ACLE_FUNC(svnbsl,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnbsl_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svnbsl_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_s32'}}
   return SVE_ACLE_FUNC(svnbsl,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnbsl_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svnbsl_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_s64'}}
   return SVE_ACLE_FUNC(svnbsl,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnbsl_n_u8u11__SVUint8_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svnbsl_n_u8(svuint8_t op1, svuint8_t op2, uint8_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_u8'}}
   return SVE_ACLE_FUNC(svnbsl,_n_u8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnbsl_n_u16u12__SVUint16_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svnbsl_n_u16(svuint16_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_u16'}}
   return SVE_ACLE_FUNC(svnbsl,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnbsl_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svnbsl_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_u32'}}
   return SVE_ACLE_FUNC(svnbsl,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svnbsl_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnbsl_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svnbsl_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svnbsl_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnbsl'}}
   // expected-warning at +1 {{implicit declaration of function 'svnbsl_n_u64'}}
   return SVE_ACLE_FUNC(svnbsl,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nmatch.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nmatch.c
index 2580268bf1518..9be3f5d0e92fe 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nmatch.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_nmatch.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,45 +15,77 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svnmatch_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nmatch.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnmatch_s8u10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nmatch.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svnmatch_s8(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svnmatch_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nmatch.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svnmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svnmatch_s8'}}
   return SVE_ACLE_FUNC(svnmatch,_s8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svnmatch_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.nmatch.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmatch_s16u10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.nmatch.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svnmatch_s16(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svnmatch_s16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.nmatch.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[RET:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[RET]]
   // overload-warning at +2 {{implicit declaration of function 'svnmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svnmatch_s16'}}
   return SVE_ACLE_FUNC(svnmatch,_s16,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svnmatch_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nmatch.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svnmatch_u8u10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nmatch.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svnmatch_u8(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svnmatch_u8
-  // CHECK: %[[intrinsic:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.nmatch.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[intrinsic]]
   // overload-warning at +2 {{implicit declaration of function 'svnmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svnmatch_u8'}}
   return SVE_ACLE_FUNC(svnmatch,_u8,,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svnmatch_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.nmatch.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svnmatch_u16u10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.nmatch.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP2]]
+//
 svbool_t test_svnmatch_u16(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svnmatch_u16
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.nmatch.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: %[[RET:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[RET]]
   // overload-warning at +2 {{implicit declaration of function 'svnmatch'}}
   // expected-warning at +1 {{implicit declaration of function 'svnmatch_u16'}}
   return SVE_ACLE_FUNC(svnmatch,_u16,,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmul.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmul.c
index c0f9b4e5f31c4..9e85b316b1208 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmul.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmul.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,22 +15,37 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpmul_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmul.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svpmul_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmul.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svpmul_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svpmul_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmul.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmul'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmul_u8'}}
   return SVE_ACLE_FUNC(svpmul,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmul_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmul.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svpmul_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmul.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svpmul_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svpmul_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmul.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmul'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmul_n_u8'}}
   return SVE_ACLE_FUNC(svpmul,_n_u8,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb.c
index c7be4eac2d751..767f4ff1fdb63 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,89 +15,153 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpmullb_pair_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svpmullb_pair_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svpmullb_pair_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_pair_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_pair_u8'}}
   return SVE_ACLE_FUNC(svpmullb_pair,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_pair_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svpmullb_pair_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svpmullb_pair_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_pair_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_pair_u32'}}
   return SVE_ACLE_FUNC(svpmullb_pair,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_pair_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svpmullb_pair_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svpmullb_pair_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_pair_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_pair_n_u8'}}
   return SVE_ACLE_FUNC(svpmullb_pair,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_pair_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svpmullb_pair_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svpmullb_pair_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_pair_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_pair_n_u32'}}
   return SVE_ACLE_FUNC(svpmullb_pair,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 16 x i8> [[TMP0]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svpmullb_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 16 x i8> [[TMP0]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svpmullb_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 16 x i8> %[[INTRINSIC]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_u16'}}
   return SVE_ACLE_FUNC(svpmullb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 4 x i32> [[TMP0]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svpmullb_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 4 x i32> [[TMP0]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svpmullb_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 4 x i32> %[[INTRINSIC]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_u64'}}
   return SVE_ACLE_FUNC(svpmullb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 16 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svpmullb_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 16 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svpmullb_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullb.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 16 x i8> %[[INTRINSIC]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_n_u16'}}
   return SVE_ACLE_FUNC(svpmullb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 4 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svpmullb_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 4 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svpmullb_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullb.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 4 x i32> %[[INTRINSIC]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_n_u64'}}
   return SVE_ACLE_FUNC(svpmullb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb_128.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb_128.c
index 41bd143e11a57..43b3f0d271669 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb_128.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullb_128.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,22 +15,37 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpmullb_pair_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullb.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svpmullb_pair_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullb.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svpmullb_pair_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_pair_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullb.pair.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_pair_u64'}}
   return SVE_ACLE_FUNC(svpmullb_pair,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullb_pair_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullb.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svpmullb_pair_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullb.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svpmullb_pair_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svpmullb_pair_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullb.pair.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullb_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullb_pair_n_u64'}}
   return SVE_ACLE_FUNC(svpmullb_pair,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt.c
index e196e9bf0dbbd..355a6ddafe7ac 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,89 +15,153 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpmullt_pair_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svpmullt_pair_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svpmullt_pair_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_pair_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_pair_u8'}}
   return SVE_ACLE_FUNC(svpmullt_pair,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_pair_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svpmullt_pair_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svpmullt_pair_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_pair_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_pair_u32'}}
   return SVE_ACLE_FUNC(svpmullt_pair,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_pair_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svpmullt_pair_n_u8u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svpmullt_pair_n_u8(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_pair_n_u8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_pair_n_u8'}}
   return SVE_ACLE_FUNC(svpmullt_pair,_n_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_pair_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svpmullt_pair_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svpmullt_pair_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_pair_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_pair_n_u32'}}
   return SVE_ACLE_FUNC(svpmullt_pair,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 16 x i8> [[TMP0]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svpmullt_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 16 x i8> [[TMP0]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svpmullt_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 16 x i8> %[[INTRINSIC]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_u16'}}
   return SVE_ACLE_FUNC(svpmullt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 4 x i32> [[TMP0]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svpmullt_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = bitcast <vscale x 4 x i32> [[TMP0]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svpmullt_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 4 x i32> %[[INTRINSIC]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_u64'}}
   return SVE_ACLE_FUNC(svpmullt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 16 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svpmullt_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 16 x i8> [[TMP1]] to <vscale x 8 x i16>
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svpmullt_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.pmullt.pair.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 16 x i8> %[[INTRINSIC]] to <vscale x 8 x i16>
-  // CHECK: ret <vscale x 8 x i16> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_n_u16'}}
   return SVE_ACLE_FUNC(svpmullt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 4 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svpmullt_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = bitcast <vscale x 4 x i32> [[TMP1]] to <vscale x 2 x i64>
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svpmullt_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.pmullt.pair.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: %[[BITCAST:.*]] = bitcast <vscale x 4 x i32> %[[INTRINSIC]] to <vscale x 2 x i64>
-  // CHECK: ret <vscale x 2 x i64> %[[BITCAST]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_n_u64'}}
   return SVE_ACLE_FUNC(svpmullt,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt_128.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt_128.c
index 398e876ab1839..298073adb3217 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt_128.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_pmullt_128.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-aes -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,22 +15,37 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svpmullt_pair_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullt.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svpmullt_pair_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullt.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svpmullt_pair_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_pair_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullt.pair.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_pair_u64'}}
   return SVE_ACLE_FUNC(svpmullt_pair,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svpmullt_pair_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullt.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svpmullt_pair_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullt.pair.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svpmullt_pair_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svpmullt_pair_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.pmullt.pair.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svpmullt_pair'}}
   // expected-warning at +1 {{implicit declaration of function 'svpmullt_pair_n_u64'}}
   return SVE_ACLE_FUNC(svpmullt_pair,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qabs.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qabs.c
index 97d6c4c95940d..4410ee505e4b0 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qabs.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qabs.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,130 +15,223 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqabs_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqabs_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqabs_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svqabs_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s8_z'}}
   return SVE_ACLE_FUNC(svqabs,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqabs_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svqabs_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s16_z'}}
   return SVE_ACLE_FUNC(svqabs,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqabs_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svqabs_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s32_z'}}
   return SVE_ACLE_FUNC(svqabs,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqabs_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svqabs_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s64_z'}}
   return SVE_ACLE_FUNC(svqabs,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqabs_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqabs_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svqabs_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s8_m'}}
   return SVE_ACLE_FUNC(svqabs,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqabs_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svqabs_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s16_m'}}
   return SVE_ACLE_FUNC(svqabs,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqabs_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svqabs_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s32_m'}}
   return SVE_ACLE_FUNC(svqabs,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqabs_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svqabs_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s64_m'}}
   return SVE_ACLE_FUNC(svqabs,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqabs_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqabs_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svqabs_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqabs.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s8_x'}}
   return SVE_ACLE_FUNC(svqabs,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqabs_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svqabs_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqabs.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s16_x'}}
   return SVE_ACLE_FUNC(svqabs,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqabs_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svqabs_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqabs.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s32_x'}}
   return SVE_ACLE_FUNC(svqabs,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqabs_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqabs_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqabs_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svqabs_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqabs.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqabs_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqabs_s64_x'}}
   return SVE_ACLE_FUNC(svqabs,_s64,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qadd.c
index 3d843dcf642f3..b6f35b2e7ff14 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,556 +15,972 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqadd_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqadd_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s8_m'}}
   return SVE_ACLE_FUNC(svqadd,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqadd_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s16_m'}}
   return SVE_ACLE_FUNC(svqadd,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqadd_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s32_m'}}
   return SVE_ACLE_FUNC(svqadd,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqadd_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s64_m'}}
   return SVE_ACLE_FUNC(svqadd,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqadd_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u8_m'}}
   return SVE_ACLE_FUNC(svqadd,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqadd_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u16_m'}}
   return SVE_ACLE_FUNC(svqadd,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqadd_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svqadd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u32_m'}}
   return SVE_ACLE_FUNC(svqadd,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqadd_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u64_m'}}
   return SVE_ACLE_FUNC(svqadd,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqadd_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqadd_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s8_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqadd_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s16_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqadd_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s32_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqadd_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s64_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqadd_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqadd_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u8_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqadd_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u16_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqadd_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u32_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqadd_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u64_m'}}
   return SVE_ACLE_FUNC(svqadd,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqadd_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s8_z'}}
   return SVE_ACLE_FUNC(svqadd,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqadd_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s16_z'}}
   return SVE_ACLE_FUNC(svqadd,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqadd_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s32_z'}}
   return SVE_ACLE_FUNC(svqadd,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqadd_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s64_z'}}
   return SVE_ACLE_FUNC(svqadd,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqadd_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u8_z'}}
   return SVE_ACLE_FUNC(svqadd,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqadd_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u16_z'}}
   return SVE_ACLE_FUNC(svqadd,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqadd_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svqadd_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u32_z'}}
   return SVE_ACLE_FUNC(svqadd,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqadd_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u64_z'}}
   return SVE_ACLE_FUNC(svqadd,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqadd_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svqadd_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s8_z
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s8_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svqadd_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s16_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svqadd_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s32_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svqadd_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s64_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqadd_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svqadd_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u8_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svqadd_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u16_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svqadd_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u32_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svqadd_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u64_z'}}
   return SVE_ACLE_FUNC(svqadd,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqadd_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s8_x'}}
   return SVE_ACLE_FUNC(svqadd,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqadd_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s16_x'}}
   return SVE_ACLE_FUNC(svqadd,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqadd_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s32_x'}}
   return SVE_ACLE_FUNC(svqadd,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqadd_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_s64_x'}}
   return SVE_ACLE_FUNC(svqadd,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqadd_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqadd_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u8_x'}}
   return SVE_ACLE_FUNC(svqadd,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqadd_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u16_x'}}
   return SVE_ACLE_FUNC(svqadd,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqadd_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svqadd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u32_x'}}
   return SVE_ACLE_FUNC(svqadd,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqadd_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqadd_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_u64_x'}}
   return SVE_ACLE_FUNC(svqadd,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqadd_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqadd_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s8_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqadd_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s16_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqadd_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s32_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqadd_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_s64_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqadd_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqadd_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u8_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqadd_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u16_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqadd_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u32_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqadd_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqadd_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqadd_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqadd_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqadd_n_u64_x'}}
   return SVE_ACLE_FUNC(svqadd,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qcadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qcadd.c
index d5956427e3634..6825960f85356 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qcadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qcadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,81 +15,137 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqcadd_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqcadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svqcadd_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqcadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqcadd_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqcadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 90)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s8'}}
   return SVE_ACLE_FUNC(svqcadd,_s8,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svqcadd_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqcadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqcadd_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqcadd.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqcadd_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqcadd.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 270)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s8'}}
   return SVE_ACLE_FUNC(svqcadd,_s8,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svqcadd_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqcadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqcadd_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqcadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqcadd_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqcadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s16'}}
   return SVE_ACLE_FUNC(svqcadd,_s16,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svqcadd_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqcadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqcadd_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqcadd.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqcadd_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqcadd.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 270)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s16'}}
   return SVE_ACLE_FUNC(svqcadd,_s16,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svqcadd_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqcadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqcadd_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqcadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqcadd_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqcadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s32'}}
   return SVE_ACLE_FUNC(svqcadd,_s32,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svqcadd_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqcadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqcadd_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqcadd.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqcadd_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqcadd.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s32'}}
   return SVE_ACLE_FUNC(svqcadd,_s32,,)(op1, op2, 270);
 }
 
+// CHECK-LABEL: @test_svqcadd_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqcadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqcadd_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqcadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqcadd_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqcadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s64'}}
   return SVE_ACLE_FUNC(svqcadd,_s64,,)(op1, op2, 90);
 }
 
+// CHECK-LABEL: @test_svqcadd_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqcadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqcadd_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqcadd.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqcadd_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqcadd_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqcadd.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqcadd'}}
   // expected-warning at +1 {{implicit declaration of function 'svqcadd_s64'}}
   return SVE_ACLE_FUNC(svqcadd,_s64,,)(op1, op2, 270);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalb.c
index 51073ec47c9ef..b9b5e3cc23d14 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,104 +15,177 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmlalb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlalb_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmlalb_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_s16'}}
   return SVE_ACLE_FUNC(svqdmlalb,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlalb_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalb_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_s32'}}
   return SVE_ACLE_FUNC(svqdmlalb,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlalb_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalb_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_s64'}}
   return SVE_ACLE_FUNC(svqdmlalb,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlalb_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmlalb_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_n_s16'}}
   return SVE_ACLE_FUNC(svqdmlalb,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlalb_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmlalb_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_n_s32'}}
   return SVE_ACLE_FUNC(svqdmlalb,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlalb_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmlalb_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_n_s64'}}
   return SVE_ACLE_FUNC(svqdmlalb,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlalb_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalb_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlalb_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlalb_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalb_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlalb_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlalb_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalb_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlalb_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlalb_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlalb_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalb_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalb_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalb_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlalb_lane,_s64,,)(op1, op2, op3, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalbt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalbt.c
index 59082de1c3092..8d323d689706a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalbt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalbt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,64 +15,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmlalbt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmlalbt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmlalbt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalbt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalbt_s16'}}
   return SVE_ACLE_FUNC(svqdmlalbt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalbt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmlalbt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalbt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalbt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalbt_s32'}}
   return SVE_ACLE_FUNC(svqdmlalbt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalbt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmlalbt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalbt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalbt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalbt_s64'}}
   return SVE_ACLE_FUNC(svqdmlalbt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalbt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdmlalbt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmlalbt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalbt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalbt_n_s16'}}
   return SVE_ACLE_FUNC(svqdmlalbt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalbt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdmlalbt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmlalbt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalbt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalbt_n_s32'}}
   return SVE_ACLE_FUNC(svqdmlalbt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalbt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdmlalbt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmlalbt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalbt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalbt_n_s64'}}
   return SVE_ACLE_FUNC(svqdmlalbt,_n_s64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalt.c
index 605bce53b0916..7a1691303e830 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlalt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,104 +15,177 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmlalt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlalt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmlalt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_s16'}}
   return SVE_ACLE_FUNC(svqdmlalt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlalt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_s32'}}
   return SVE_ACLE_FUNC(svqdmlalt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlalt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_s64'}}
   return SVE_ACLE_FUNC(svqdmlalt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlalt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmlalt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlalt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_n_s16'}}
   return SVE_ACLE_FUNC(svqdmlalt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlalt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmlalt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_n_s32'}}
   return SVE_ACLE_FUNC(svqdmlalt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlalt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmlalt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_n_s64'}}
   return SVE_ACLE_FUNC(svqdmlalt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlalt_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalt_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlalt_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlalt_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlalt_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlalt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlalt_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlalt_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalt_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlalt_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlalt_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlalt_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlalt_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlalt_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlalt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlalt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlalt_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlalt_lane,_s64,,)(op1, op2, op3, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslb.c
index f331e6f24d1fc..e4fd60422eba2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,104 +15,177 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmlslb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlslb_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmlslb_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_s16'}}
   return SVE_ACLE_FUNC(svqdmlslb,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlslb_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslb_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_s32'}}
   return SVE_ACLE_FUNC(svqdmlslb,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlslb_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslb_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_s64'}}
   return SVE_ACLE_FUNC(svqdmlslb,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlslb_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmlslb_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_n_s16'}}
   return SVE_ACLE_FUNC(svqdmlslb,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlslb_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmlslb_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_n_s32'}}
   return SVE_ACLE_FUNC(svqdmlslb,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlslb_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmlslb_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_n_s64'}}
   return SVE_ACLE_FUNC(svqdmlslb,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlslb_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslb_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlslb_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlslb_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslb_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslb.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlslb_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlslb_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslb_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlslb_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlslb_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlslb_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslb_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslb_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslb.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslb_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlslb_lane,_s64,,)(op1, op2, op3, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslbt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslbt.c
index 4584e4ff5d7e4..56a5fe1a1788e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslbt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslbt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,64 +15,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmlslbt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmlslbt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmlslbt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslbt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslbt_s16'}}
   return SVE_ACLE_FUNC(svqdmlslbt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslbt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmlslbt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslbt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslbt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslbt_s32'}}
   return SVE_ACLE_FUNC(svqdmlslbt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslbt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmlslbt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslbt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslbt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslbt_s64'}}
   return SVE_ACLE_FUNC(svqdmlslbt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslbt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdmlslbt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslbt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmlslbt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslbt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslbt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslbt_n_s16'}}
   return SVE_ACLE_FUNC(svqdmlslbt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslbt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdmlslbt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslbt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmlslbt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslbt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslbt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslbt_n_s32'}}
   return SVE_ACLE_FUNC(svqdmlslbt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslbt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqdmlslbt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslbt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmlslbt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslbt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslbt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslbt_n_s64'}}
   return SVE_ACLE_FUNC(svqdmlslbt,_n_s64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslt.c
index 52694173d2a9d..30444905ac21c 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmlslt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,104 +15,177 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmlslt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlslt_s16u11__SVInt16_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmlslt_s16(svint16_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_s16'}}
   return SVE_ACLE_FUNC(svqdmlslt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlslt_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslt_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_s32'}}
   return SVE_ACLE_FUNC(svqdmlslt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmlslt_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslt_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_s64'}}
   return SVE_ACLE_FUNC(svqdmlslt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlslt_n_s16u11__SVInt16_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmlslt_n_s16(svint16_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmlslt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_n_s16'}}
   return SVE_ACLE_FUNC(svqdmlslt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlslt_n_s32u11__SVInt32_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmlslt_n_s32(svint32_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_n_s32'}}
   return SVE_ACLE_FUNC(svqdmlslt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmlslt_n_s64u11__SVInt64_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmlslt_n_s64(svint64_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_n_s64'}}
   return SVE_ACLE_FUNC(svqdmlslt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlslt_lane_s32u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslt_lane_s32(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlslt_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlslt_lane_s32_1u11__SVInt32_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmlslt_lane_s32_1(svint32_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmlslt.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmlslt_lane,_s32,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmlslt_lane_s64u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslt_lane_s64(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlslt_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqdmlslt_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmlslt_lane_s64_1u11__SVInt64_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmlslt_lane_s64_1(svint64_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqdmlslt_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmlslt.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmlslt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmlslt_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmlslt_lane,_s64,,)(op1, op2, op3, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmulh.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmulh.c
index 7d48bca9c7be3..ad3279d67ecc0 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmulh.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmulh.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,145 +15,247 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmulh_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqdmulh_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqdmulh_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqdmulh.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_s8'}}
   return SVE_ACLE_FUNC(svqdmulh,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqdmulh_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmulh_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_s16'}}
   return SVE_ACLE_FUNC(svqdmulh,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqdmulh_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmulh_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_s32'}}
   return SVE_ACLE_FUNC(svqdmulh,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqdmulh_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmulh_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_s64'}}
   return SVE_ACLE_FUNC(svqdmulh,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmulh_n_s8u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqdmulh_n_s8(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqdmulh.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_n_s8'}}
   return SVE_ACLE_FUNC(svqdmulh,_n_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmulh_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmulh_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_n_s16'}}
   return SVE_ACLE_FUNC(svqdmulh,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmulh_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmulh_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_n_s32'}}
   return SVE_ACLE_FUNC(svqdmulh,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqdmulh_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmulh_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_n_s64'}}
   return SVE_ACLE_FUNC(svqdmulh,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmulh_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdmulh_lane_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmulh_lane_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_lane_s16'}}
   return SVE_ACLE_FUNC(svqdmulh_lane,_s16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmulh_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svqdmulh_lane_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmulh_lane_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmulh.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_lane_s16'}}
   return SVE_ACLE_FUNC(svqdmulh_lane,_s16,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svqdmulh_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdmulh_lane_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmulh_lane_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmulh_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmulh_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svqdmulh_lane_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmulh_lane_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmulh.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmulh_lane,_s32,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svqdmulh_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqdmulh_lane_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmulh_lane_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmulh_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmulh_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svqdmulh_lane_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmulh_lane_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqdmulh_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmulh.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmulh_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmulh_lane,_s64,,)(op1, op2, 1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullb.c
index 62fb6ecde66fd..61fedbfc0bb35 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,104 +15,177 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmullb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmullb_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmullb_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_s16'}}
   return SVE_ACLE_FUNC(svqdmullb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmullb_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmullb_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_s32'}}
   return SVE_ACLE_FUNC(svqdmullb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmullb_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmullb_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_s64'}}
   return SVE_ACLE_FUNC(svqdmullb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmullb_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmullb_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_n_s16'}}
   return SVE_ACLE_FUNC(svqdmullb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmullb_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmullb_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_n_s32'}}
   return SVE_ACLE_FUNC(svqdmullb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmullb_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmullb_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_n_s64'}}
   return SVE_ACLE_FUNC(svqdmullb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullb_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmullb_lane_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmullb_lane_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmullb_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmullb_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmullb_lane_s32_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmullb_lane_s32_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullb.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmullb_lane,_s32,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svqdmullb_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmullb_lane_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmullb_lane_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmullb_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmullb_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmullb_lane_s64_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmullb_lane_s64_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullb_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullb.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullb_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullb_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmullb_lane,_s64,,)(op1, op2, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullt.c
index df4262e548747..d14c758274cd9 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qdmullt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,104 +15,177 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqdmullt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmullt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqdmullt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_s16'}}
   return SVE_ACLE_FUNC(svqdmullt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmullt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmullt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_s32'}}
   return SVE_ACLE_FUNC(svqdmullt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqdmullt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmullt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_s64'}}
   return SVE_ACLE_FUNC(svqdmullt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmullt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqdmullt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqdmullt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_n_s16'}}
   return SVE_ACLE_FUNC(svqdmullt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmullt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqdmullt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_n_s32'}}
   return SVE_ACLE_FUNC(svqdmullt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqdmullt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqdmullt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_n_s64'}}
   return SVE_ACLE_FUNC(svqdmullt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqdmullt_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmullt_lane_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmullt_lane_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmullt_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmullt_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmullt_lane_s32_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.lane.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqdmullt_lane_s32_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqdmullt.lane.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_lane_s32'}}
   return SVE_ACLE_FUNC(svqdmullt_lane,_s32,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svqdmullt_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqdmullt_lane_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmullt_lane_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmullt_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqdmullt_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqdmullt_lane_s64_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.lane.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqdmullt_lane_s64_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqdmullt_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqdmullt.lane.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqdmullt_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqdmullt_lane_s64'}}
   return SVE_ACLE_FUNC(svqdmullt_lane,_s64,,)(op1, op2, 3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qneg.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qneg.c
index 59af8c87fc69d..f07f88a595551 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qneg.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qneg.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,130 +15,223 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqneg_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqneg_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqneg_s8_z(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svqneg_s8_z
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> zeroinitializer, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s8_z'}}
   return SVE_ACLE_FUNC(svqneg,_s8,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqneg_s16_z(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svqneg_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> zeroinitializer, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s16_z'}}
   return SVE_ACLE_FUNC(svqneg,_s16,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqneg_s32_z(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svqneg_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s32_z'}}
   return SVE_ACLE_FUNC(svqneg,_s32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqneg_s64_z(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svqneg_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> zeroinitializer, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s64_z'}}
   return SVE_ACLE_FUNC(svqneg,_s64,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqneg_s8_mu10__SVInt8_tu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> [[INACTIVE:%.*]], <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqneg_s8_m(svint8_t inactive, svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svqneg_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> %inactive, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s8_m'}}
   return SVE_ACLE_FUNC(svqneg,_s8,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s16_mu11__SVInt16_tu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> [[INACTIVE:%.*]], <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqneg_s16_m(svint16_t inactive, svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svqneg_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> %inactive, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s16_m'}}
   return SVE_ACLE_FUNC(svqneg,_s16,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s32_mu11__SVInt32_tu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqneg_s32_m(svint32_t inactive, svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svqneg_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s32_m'}}
   return SVE_ACLE_FUNC(svqneg,_s32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s64_mu11__SVInt64_tu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> [[INACTIVE:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqneg_s64_m(svint64_t inactive, svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svqneg_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> %inactive, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s64_m'}}
   return SVE_ACLE_FUNC(svqneg,_s64,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqneg_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqneg_s8_x(svbool_t pg, svint8_t op)
 {
-  // CHECK-LABEL: test_svqneg_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqneg.nxv16i8(<vscale x 16 x i8> undef, <vscale x 16 x i1> %pg, <vscale x 16 x i8> %op)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s8_x'}}
   return SVE_ACLE_FUNC(svqneg,_s8,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqneg_s16_x(svbool_t pg, svint16_t op)
 {
-  // CHECK-LABEL: test_svqneg_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqneg.nxv8i16(<vscale x 8 x i16> undef, <vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s16_x'}}
   return SVE_ACLE_FUNC(svqneg,_s16,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqneg_s32_x(svbool_t pg, svint32_t op)
 {
-  // CHECK-LABEL: test_svqneg_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqneg.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s32_x'}}
   return SVE_ACLE_FUNC(svqneg,_s32,_x,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svqneg_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqneg_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqneg_s64_x(svbool_t pg, svint64_t op)
 {
-  // CHECK-LABEL: test_svqneg_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqneg.nxv2i64(<vscale x 2 x i64> undef, <vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqneg_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqneg_s64_x'}}
   return SVE_ACLE_FUNC(svqneg,_s64,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdcmlah.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdcmlah.c
index 4b8dd93bf08b2..471968a380a28 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdcmlah.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdcmlah.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,201 +15,341 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrdcmlah_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdcmlah_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdcmlah_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s8'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s8,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdcmlah_s8_1u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdcmlah_s8_1(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 90)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s8'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s8,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s8_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdcmlah_s8_2u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdcmlah_s8_2(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s8_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 180)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s8'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s8,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s8_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdcmlah_s8_3u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdcmlah_s8_3(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s8_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdcmlah.x.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3, i32 270)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s8'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s8,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrdcmlah_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdcmlah_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s16'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdcmlah_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s16'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s16,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s16_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s16_2u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdcmlah_s16_2(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s16_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 180)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s16'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s16,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s16_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s16_3u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdcmlah_s16_3(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s16_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 270)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s16'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s16,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrdcmlah_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdcmlah_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s32'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdcmlah_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 90)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s32'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s32,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s32_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s32_2u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdcmlah_s32_2(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s32_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 180)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s32'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s32,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s32_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s32_3u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdcmlah_s32_3(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s32_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s32'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s32,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrdcmlah_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdcmlah_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s64'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 90)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s64_1u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdcmlah_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 90)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s64'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s64,,)(op1, op2, op3, 90);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s64_2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 180)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s64_2u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdcmlah_s64_2(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s64_2
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 180)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s64'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s64,,)(op1, op2, op3, 180);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_s64_3(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 270)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrdcmlah_s64_3u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdcmlah_s64_3(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_s64_3
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdcmlah.x.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 270)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_s64'}}
   return SVE_ACLE_FUNC(svqrdcmlah,_s64,,)(op1, op2, op3, 270);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svqrdcmlah_lane_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0, i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdcmlah_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdcmlah_lane,_s16,,)(op1, op2, op3, 0, 0);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3, i32 90)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svqrdcmlah_lane_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 3, i32 90)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdcmlah_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 3, i32 90)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdcmlah_lane,_s16,,)(op1, op2, op3, 3, 90);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0, i32 180)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z24test_svqrdcmlah_lane_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0, i32 180)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdcmlah_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0, i32 180)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdcmlah_lane,_s32,,)(op1, op2, op3, 0, 180);
 }
 
+// CHECK-LABEL: @test_svqrdcmlah_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 1, i32 270)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z26test_svqrdcmlah_lane_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 1, i32 270)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdcmlah_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdcmlah_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdcmlah.lane.x.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 1, i32 270)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdcmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdcmlah_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdcmlah_lane,_s32,,)(op1, op2, op3, 1, 270);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlah.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlah.c
index 5cc49a713072a..af000b6284024 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlah.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlah.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,145 +15,247 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrdmlah_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlah.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrdmlah_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlah.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdmlah_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlah.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_s8'}}
   return SVE_ACLE_FUNC(svqrdmlah,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmlah_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmlah_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_s16'}}
   return SVE_ACLE_FUNC(svqrdmlah,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmlah_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmlah_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_s32'}}
   return SVE_ACLE_FUNC(svqrdmlah,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmlah_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmlah_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_s64'}}
   return SVE_ACLE_FUNC(svqrdmlah,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlah.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrdmlah_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlah.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqrdmlah_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlah.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_n_s8'}}
   return SVE_ACLE_FUNC(svqrdmlah,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmlah_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqrdmlah_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_n_s16'}}
   return SVE_ACLE_FUNC(svqrdmlah,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmlah_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqrdmlah_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_n_s32'}}
   return SVE_ACLE_FUNC(svqrdmlah,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmlah_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqrdmlah_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_n_s64'}}
   return SVE_ACLE_FUNC(svqrdmlah,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmlah_lane_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmlah_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdmlah_lane,_s16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmlah_lane_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmlah_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlah.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdmlah_lane,_s16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmlah_lane_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmlah_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdmlah_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmlah_lane_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmlah_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlah.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdmlah_lane,_s32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmlah_lane_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmlah_lane_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_lane_s64'}}
   return SVE_ACLE_FUNC(svqrdmlah_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmlah_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmlah_lane_s64_1u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmlah_lane_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlah_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlah.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlah_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlah_lane_s64'}}
   return SVE_ACLE_FUNC(svqrdmlah_lane,_s64,,)(op1, op2, op3, 1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlsh.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlsh.c
index 3838bdc53e1ef..07b36721bf3b4 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlsh.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmlsh.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,145 +15,247 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrdmlsh_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlsh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrdmlsh_s8u10__SVInt8_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlsh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdmlsh_s8(svint8_t op1, svint8_t op2, svint8_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlsh.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_s8'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmlsh_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmlsh_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_s16'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmlsh_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmlsh_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_s32'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmlsh_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmlsh_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_s64'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlsh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrdmlsh_n_s8u10__SVInt8_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlsh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqrdmlsh_n_s8(svint8_t op1, svint8_t op2, int8_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmlsh.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_n_s8'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_n_s8,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmlsh_n_s16u11__SVInt16_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqrdmlsh_n_s16(svint16_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_n_s16'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmlsh_n_s32u11__SVInt32_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqrdmlsh_n_s32(svint32_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_n_s32'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmlsh_n_s64u11__SVInt64_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqrdmlsh_n_s64(svint64_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_n_s64'}}
   return SVE_ACLE_FUNC(svqrdmlsh,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmlsh_lane_s16u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmlsh_lane_s16(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdmlsh_lane,_s16,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmlsh_lane_s16_1u11__SVInt16_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmlsh_lane_s16_1(svint16_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmlsh.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdmlsh_lane,_s16,,)(op1, op2, op3, 7);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmlsh_lane_s32u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmlsh_lane_s32(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdmlsh_lane,_s32,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmlsh_lane_s32_1u11__SVInt32_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmlsh_lane_s32_1(svint32_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmlsh.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdmlsh_lane,_s32,,)(op1, op2, op3, 3);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmlsh_lane_s64u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmlsh_lane_s64(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_lane_s64'}}
   return SVE_ACLE_FUNC(svqrdmlsh_lane,_s64,,)(op1, op2, op3, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmlsh_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmlsh_lane_s64_1u11__SVInt64_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmlsh_lane_s64_1(svint64_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svqrdmlsh_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmlsh.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmlsh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmlsh_lane_s64'}}
   return SVE_ACLE_FUNC(svqrdmlsh_lane,_s64,,)(op1, op2, op3, 1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmulh.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmulh.c
index e009441ffcbfd..0a5727d8599f2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmulh.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrdmulh.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,145 +15,247 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrdmulh_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrdmulh_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrdmulh_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmulh.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_s8'}}
   return SVE_ACLE_FUNC(svqrdmulh,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmulh_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmulh_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_s16'}}
   return SVE_ACLE_FUNC(svqrdmulh,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmulh_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmulh_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_s32'}}
   return SVE_ACLE_FUNC(svqrdmulh,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrdmulh_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmulh_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_s64'}}
   return SVE_ACLE_FUNC(svqrdmulh,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrdmulh_n_s8u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmulh.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqrdmulh_n_s8(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_n_s8
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrdmulh.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_n_s8'}}
   return SVE_ACLE_FUNC(svqrdmulh,_n_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmulh_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqrdmulh_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_n_s16'}}
   return SVE_ACLE_FUNC(svqrdmulh,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmulh_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqrdmulh_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_n_s32'}}
   return SVE_ACLE_FUNC(svqrdmulh,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrdmulh_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqrdmulh_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_n_s64'}}
   return SVE_ACLE_FUNC(svqrdmulh,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_lane_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmulh_lane_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmulh_lane_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_lane_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdmulh_lane,_s16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_lane_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmulh_lane_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.lane.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrdmulh_lane_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_lane_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrdmulh.lane.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_lane_s16'}}
   return SVE_ACLE_FUNC(svqrdmulh_lane,_s16,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_lane_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmulh_lane_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmulh_lane_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_lane_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdmulh_lane,_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_lane_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmulh_lane_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.lane.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 3)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrdmulh_lane_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_lane_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrdmulh.lane.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_lane_s32'}}
   return SVE_ACLE_FUNC(svqrdmulh_lane,_s32,,)(op1, op2, 3);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_lane_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrdmulh_lane_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmulh_lane_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_lane_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_lane_s64'}}
   return SVE_ACLE_FUNC(svqrdmulh_lane,_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svqrdmulh_lane_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z25test_svqrdmulh_lane_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.lane.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svqrdmulh_lane_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrdmulh_lane_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrdmulh.lane.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrdmulh_lane'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrdmulh_lane_s64'}}
   return SVE_ACLE_FUNC(svqrdmulh_lane,_s64,,)(op1, op2, 1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshl.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshl.c
index 475e2ef221176..38c6dbf5816eb 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshl.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshl.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrshl_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrshl_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqrshl_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s8_z'}}
   return SVE_ACLE_FUNC(svqrshl,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqrshl_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s16_z'}}
   return SVE_ACLE_FUNC(svqrshl,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqrshl_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s32_z'}}
   return SVE_ACLE_FUNC(svqrshl,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqrshl_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s64_z'}}
   return SVE_ACLE_FUNC(svqrshl,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrshl_u8_zu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqrshl_u8_z(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u8_z'}}
   return SVE_ACLE_FUNC(svqrshl,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u16_zu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqrshl_u16_z(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u16_z'}}
   return SVE_ACLE_FUNC(svqrshl,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u32_zu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqrshl_u32_z(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u32_z'}}
   return SVE_ACLE_FUNC(svqrshl,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u64_zu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqrshl_u64_z(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u64_z'}}
   return SVE_ACLE_FUNC(svqrshl,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrshl_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrshl_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s8_m'}}
   return SVE_ACLE_FUNC(svqrshl,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqrshl_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s16_m'}}
   return SVE_ACLE_FUNC(svqrshl,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqrshl_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s32_m'}}
   return SVE_ACLE_FUNC(svqrshl,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqrshl_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s64_m'}}
   return SVE_ACLE_FUNC(svqrshl,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrshl_u8_mu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshl_u8_m(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u8_m'}}
   return SVE_ACLE_FUNC(svqrshl,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u16_mu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqrshl_u16_m(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u16_m'}}
   return SVE_ACLE_FUNC(svqrshl,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u32_mu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqrshl_u32_m(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u32_m'}}
   return SVE_ACLE_FUNC(svqrshl,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u64_mu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqrshl_u64_m(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u64_m'}}
   return SVE_ACLE_FUNC(svqrshl,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrshl_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrshl_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s8_x'}}
   return SVE_ACLE_FUNC(svqrshl,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqrshl_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s16_x'}}
   return SVE_ACLE_FUNC(svqrshl,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqrshl_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s32_x'}}
   return SVE_ACLE_FUNC(svqrshl,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqrshl_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_s64_x'}}
   return SVE_ACLE_FUNC(svqrshl,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqrshl_u8_xu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshl_u8_x(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u8_x'}}
   return SVE_ACLE_FUNC(svqrshl,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u16_xu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqrshl_u16_x(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u16_x'}}
   return SVE_ACLE_FUNC(svqrshl,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u32_xu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqrshl_u32_x(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u32_x'}}
   return SVE_ACLE_FUNC(svqrshl,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqrshl_u64_xu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqrshl_u64_x(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_u64_x'}}
   return SVE_ACLE_FUNC(svqrshl,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrshl_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svqrshl_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s8_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svqrshl_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s16_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svqrshl_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s32_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svqrshl_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s64_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrshl_n_u8_zu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svqrshl_n_u8_z(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u8_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u16_zu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svqrshl_n_u16_z(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u16_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u32_zu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svqrshl_n_u32_z(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u32_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u64_zu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svqrshl_n_u64_z(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u64_z'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrshl_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqrshl_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s8_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqrshl_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s16_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqrshl_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s32_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqrshl_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s64_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrshl_n_u8_mu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqrshl_n_u8_m(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u8_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u16_mu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqrshl_n_u16_m(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u16_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u32_mu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqrshl_n_u32_m(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u32_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u64_mu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqrshl_n_u64_m(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u64_m'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrshl_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqrshl_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s8_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqrshl_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s16_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqrshl_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s32_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqrshl_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_s64_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqrshl_n_u8_xu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqrshl_n_u8_x(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u8_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u16_xu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqrshl_n_u16_x(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u16_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u32_xu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqrshl_n_u32_x(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u32_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqrshl_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshl_n_u64_xu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqrshl_n_u64_x(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqrshl_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqrshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshl_n_u64_x'}}
   return SVE_ACLE_FUNC(svqrshl,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnb.c
index 3e31a7ecaf252..5ddc73f0faaca 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrshrnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnb_n_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrshrnb_n_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_s16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnb_n_s16_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrshrnb_n_s16_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_s16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnb_n_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrshrnb_n_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_s32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnb_n_s32_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrshrnb_n_s32_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_s32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnb_n_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrshrnb_n_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_s64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnb_n_s64_1u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrshrnb_n_s64_1(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_s64,,)(op1, 32);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnb_n_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrnb_n_u16(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_u16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnb_n_u16_1u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrnb_n_u16_1(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_u16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnb_n_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrnb_n_u32(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_u32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnb_n_u32_1u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrnb_n_u32_1(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_u32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnb_n_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrnb_n_u64(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_u64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnb_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnb_n_u64_1u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrnb_n_u64_1(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnb_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svqrshrnb,_n_u64,,)(op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnt.c
index 0f996502d9bda..91584307ed938 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrshrnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnt_n_s16u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrshrnt_n_s16(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_s16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnt_n_s16_1u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqrshrnt_n_s16_1(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_s16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnt_n_s32u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrshrnt_n_s32(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_s32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnt_n_s32_1u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqrshrnt_n_s32_1(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_s32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnt_n_s64u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrshrnt_n_s64(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_s64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnt_n_s64_1u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqrshrnt_n_s64_1(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_s64,,)(op, op1, 32);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnt_n_u16u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrnt_n_u16(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_u16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnt_n_u16_1u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrnt_n_u16_1(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqrshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_u16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnt_n_u32u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrnt_n_u32(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_u32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnt_n_u32_1u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrnt_n_u32_1(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqrshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_u32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqrshrnt_n_u64u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrnt_n_u64(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_u64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrnt_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqrshrnt_n_u64_1u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrnt_n_u64_1(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrnt_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqrshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svqrshrnt,_n_u64,,)(op, op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunb.c
index a4256ec54c620..9d767998cd2af 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrshrunb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrshrunb_n_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrunb_n_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunb_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrunb,_n_s16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrunb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrshrunb_n_s16_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrunb_n_s16_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunb_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrunb,_n_s16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svqrshrunb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrshrunb_n_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrunb_n_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunb_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrunb,_n_s32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrunb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrshrunb_n_s32_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrunb_n_s32_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunb_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrunb,_n_s32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svqrshrunb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrshrunb_n_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrunb_n_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunb_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrunb,_n_s64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrunb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrshrunb_n_s64_1u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrunb_n_s64_1(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunb_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrunb,_n_s64,,)(op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunt.c
index 2ae6c61db07f4..bb3a90bb36be6 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qrshrunt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqrshrunt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrshrunt_n_s16u11__SVUint8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrunt_n_s16(svuint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunt_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrunt,_n_s16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrunt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrshrunt_n_s16_1u11__SVUint8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqrshrunt_n_s16_1(svuint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqrshrunt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunt_n_s16'}}
   return SVE_ACLE_FUNC(svqrshrunt,_n_s16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svqrshrunt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrshrunt_n_s32u12__SVUint16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrunt_n_s32(svuint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunt_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrunt,_n_s32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrunt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrshrunt_n_s32_1u12__SVUint16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqrshrunt_n_s32_1(svuint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqrshrunt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunt_n_s32'}}
   return SVE_ACLE_FUNC(svqrshrunt,_n_s32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svqrshrunt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqrshrunt_n_s64u12__SVUint32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrunt_n_s64(svuint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunt_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrunt,_n_s64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqrshrunt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z23test_svqrshrunt_n_s64_1u12__SVUint32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqrshrunt_n_s64_1(svuint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqrshrunt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqrshrunt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqrshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqrshrunt_n_s64'}}
   return SVE_ACLE_FUNC(svqrshrunt,_n_s64,,)(op, op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshl.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshl.c
index 23aea5df55d61..3fde6854a1b4d 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshl.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshl.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqshl_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqshl_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqshl_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s8_z'}}
   return SVE_ACLE_FUNC(svqshl,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqshl_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s16_z'}}
   return SVE_ACLE_FUNC(svqshl,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqshl_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s32_z'}}
   return SVE_ACLE_FUNC(svqshl,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqshl_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s64_z'}}
   return SVE_ACLE_FUNC(svqshl,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqshl_u8_zu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqshl_u8_z(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u8_z'}}
   return SVE_ACLE_FUNC(svqshl,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u16_zu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqshl_u16_z(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u16_z'}}
   return SVE_ACLE_FUNC(svqshl,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u32_zu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqshl_u32_z(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u32_z'}}
   return SVE_ACLE_FUNC(svqshl,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u64_zu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqshl_u64_z(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u64_z'}}
   return SVE_ACLE_FUNC(svqshl,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqshl_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqshl_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s8_m'}}
   return SVE_ACLE_FUNC(svqshl,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqshl_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s16_m'}}
   return SVE_ACLE_FUNC(svqshl,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqshl_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s32_m'}}
   return SVE_ACLE_FUNC(svqshl,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqshl_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s64_m'}}
   return SVE_ACLE_FUNC(svqshl,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqshl_u8_mu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshl_u8_m(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u8_m'}}
   return SVE_ACLE_FUNC(svqshl,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u16_mu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqshl_u16_m(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u16_m'}}
   return SVE_ACLE_FUNC(svqshl,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u32_mu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqshl_u32_m(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u32_m'}}
   return SVE_ACLE_FUNC(svqshl,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u64_mu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqshl_u64_m(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u64_m'}}
   return SVE_ACLE_FUNC(svqshl,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqshl_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqshl_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s8_x'}}
   return SVE_ACLE_FUNC(svqshl,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqshl_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s16_x'}}
   return SVE_ACLE_FUNC(svqshl,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqshl_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s32_x'}}
   return SVE_ACLE_FUNC(svqshl,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqshl_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_s64_x'}}
   return SVE_ACLE_FUNC(svqshl,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqshl_u8_xu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshl_u8_x(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u8_x'}}
   return SVE_ACLE_FUNC(svqshl,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u16_xu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqshl_u16_x(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u16_x'}}
   return SVE_ACLE_FUNC(svqshl,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u32_xu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqshl_u32_x(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u32_x'}}
   return SVE_ACLE_FUNC(svqshl,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqshl_u64_xu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqshl_u64_x(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_u64_x'}}
   return SVE_ACLE_FUNC(svqshl,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqshl_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svqshl_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s8_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svqshl_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s16_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svqshl_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s32_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svqshl_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s64_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqshl_n_u8_zu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svqshl_n_u8_z(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u8_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u16_zu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svqshl_n_u16_z(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u16_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u32_zu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svqshl_n_u32_z(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u32_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u64_zu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svqshl_n_u64_z(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u64_z'}}
   return SVE_ACLE_FUNC(svqshl,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqshl_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqshl_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s8_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqshl_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s16_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqshl_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s32_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqshl_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s64_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqshl_n_u8_mu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqshl_n_u8_m(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u8_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u16_mu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqshl_n_u16_m(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u16_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u32_mu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqshl_n_u32_m(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u32_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u64_mu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqshl_n_u64_m(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u64_m'}}
   return SVE_ACLE_FUNC(svqshl,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqshl_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqshl_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s8_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqshl_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s16_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqshl_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s32_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqshl_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_s64_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqshl_n_u8_xu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqshl_n_u8_x(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u8_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u16_xu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqshl_n_u16_x(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u16_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u32_xu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqshl_n_u32_x(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u32_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqshl_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshl_n_u64_xu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqshl_n_u64_x(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqshl_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshl_n_u64_x'}}
   return SVE_ACLE_FUNC(svqshl,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshlu.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshlu.c
index 93618ea3ceebf..0df31a9a303a4 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshlu.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshlu.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,267 +15,461 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqshlu_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshlu_n_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqshlu_n_s8_z(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s8_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s8,_z,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s8_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 7)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshlu_n_s8_z_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqshlu_n_s8_z_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s8_z_1
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 7)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s8_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s8,_z,)(pg, op1, 7);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqshlu_n_s16_z(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s16_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s16,_z,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s16_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 15)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s16_z_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqshlu_n_s16_z_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s16_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 15)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s16_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s16,_z,)(pg, op1, 15);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqshlu_n_s32_z(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s32_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s32,_z,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s32_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 31)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s32_z_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqshlu_n_s32_z_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s32_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 31)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s32_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s32,_z,)(pg, op1, 31);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqshlu_n_s64_z(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s64_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s64,_z,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s64_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 63)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s64_z_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqshlu_n_s64_z_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s64_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 63)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s64_z'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s64,_z,)(pg, op1, 63);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshlu_n_s8_mu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshlu_n_s8_m(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s8_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s8,_m,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s8_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshlu_n_s8_m_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshlu_n_s8_m_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s8_m_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 7)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s8_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s8,_m,)(pg, op1, 7);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s16_mu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqshlu_n_s16_m(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s16_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s16,_m,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s16_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s16_m_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqshlu_n_s16_m_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s16_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 15)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s16_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s16,_m,)(pg, op1, 15);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s32_mu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqshlu_n_s32_m(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s32_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s32,_m,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s32_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s32_m_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqshlu_n_s32_m_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s32_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 31)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s32_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s32,_m,)(pg, op1, 31);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s64_mu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqshlu_n_s64_m(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s64_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s64,_m,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s64_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s64_m_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqshlu_n_s64_m_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s64_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 63)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s64_m'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s64,_m,)(pg, op1, 63);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshlu_n_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshlu_n_s8_x(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s8_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s8,_x,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s8_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshlu_n_s8_x_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshlu_n_s8_x_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s8_x_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshlu.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 7)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s8_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s8,_x,)(pg, op1, 7);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqshlu_n_s16_x(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s16_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s16,_x,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s16_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s16_x_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqshlu_n_s16_x_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s16_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshlu.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 15)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s16_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s16,_x,)(pg, op1, 15);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqshlu_n_s32_x(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s32_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s32,_x,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s32_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s32_x_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqshlu_n_s32_x_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s32_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshlu.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 31)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s32_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s32,_x,)(pg, op1, 31);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshlu_n_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqshlu_n_s64_x(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s64_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s64,_x,)(pg, op1, 0);
 }
 
+// CHECK-LABEL: @test_svqshlu_n_s64_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshlu_n_s64_x_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqshlu_n_s64_x_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshlu_n_s64_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqshlu.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 63)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshlu_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshlu_n_s64_x'}}
   return SVE_ACLE_FUNC(svqshlu,_n_s64,_x,)(pg, op1, 63);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnb.c
index a06cc62a94bbf..af93d5952e469 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqshrnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnb_n_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqshrnb_n_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_s16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnb_n_s16_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqshrnb_n_s16_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_s16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnb_n_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqshrnb_n_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_s32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnb_n_s32_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqshrnb_n_s32_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_s32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnb_n_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqshrnb_n_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_s64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnb_n_s64_1u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqshrnb_n_s64_1(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_s64,,)(op1, 32);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnb_n_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrnb_n_u16(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_u16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnb_n_u16_1u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrnb_n_u16_1(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_u16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnb_n_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrnb_n_u32(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_u32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnb_n_u32_1u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrnb_n_u32_1(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_u32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnb_n_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrnb_n_u64(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_u64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnb_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnb_n_u64_1u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrnb_n_u64_1(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnb_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svqshrnb,_n_u64,,)(op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnt.c
index 13a82a36272af..fed19c2841b34 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqshrnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnt_n_s16u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqshrnt_n_s16(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_s16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnt_n_s16_1u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqshrnt_n_s16_1(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_s16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnt_n_s32u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqshrnt_n_s32(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_s32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnt_n_s32_1u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqshrnt_n_s32_1(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_s32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnt_n_s64u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqshrnt_n_s64(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_s64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnt_n_s64_1u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqshrnt_n_s64_1(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_s64,,)(op, op1, 32);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnt_n_u16u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrnt_n_u16(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_u16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnt_n_u16_1u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrnt_n_u16_1(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_u16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnt_n_u32u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrnt_n_u32(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_u32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnt_n_u32_1u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrnt_n_u32_1(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_u32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqshrnt_n_u64u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrnt_n_u64(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_u64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrnt_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svqshrnt_n_u64_1u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrnt_n_u64_1(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrnt_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svqshrnt,_n_u64,,)(op, op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunb.c
index ab218f35a83c3..e9bf259b12b9b 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqshrunb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshrunb_n_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrunb_n_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrunb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunb_n_s16'}}
   return SVE_ACLE_FUNC(svqshrunb,_n_s16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrunb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshrunb_n_s16_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrunb_n_s16_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrunb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunb_n_s16'}}
   return SVE_ACLE_FUNC(svqshrunb,_n_s16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svqshrunb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshrunb_n_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrunb_n_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrunb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunb_n_s32'}}
   return SVE_ACLE_FUNC(svqshrunb,_n_s32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrunb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshrunb_n_s32_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrunb_n_s32_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrunb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunb_n_s32'}}
   return SVE_ACLE_FUNC(svqshrunb,_n_s32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svqshrunb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshrunb_n_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrunb_n_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrunb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunb_n_s64'}}
   return SVE_ACLE_FUNC(svqshrunb,_n_s64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrunb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshrunb_n_s64_1u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrunb_n_s64_1(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrunb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunb_n_s64'}}
   return SVE_ACLE_FUNC(svqshrunb,_n_s64,,)(op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunt.c
index 7df3a4227d7c7..39912d9aca035 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qshrunt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqshrunt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshrunt_n_s16u11__SVUint8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrunt_n_s16(svuint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrunt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunt_n_s16'}}
   return SVE_ACLE_FUNC(svqshrunt,_n_s16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrunt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshrunt_n_s16_1u11__SVUint8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqshrunt_n_s16_1(svuint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqshrunt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqshrunt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunt_n_s16'}}
   return SVE_ACLE_FUNC(svqshrunt,_n_s16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svqshrunt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshrunt_n_s32u12__SVUint16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrunt_n_s32(svuint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrunt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunt_n_s32'}}
   return SVE_ACLE_FUNC(svqshrunt,_n_s32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrunt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshrunt_n_s32_1u12__SVUint16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqshrunt_n_s32_1(svuint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqshrunt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqshrunt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunt_n_s32'}}
   return SVE_ACLE_FUNC(svqshrunt,_n_s32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svqshrunt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqshrunt_n_s64u12__SVUint32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrunt_n_s64(svuint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrunt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunt_n_s64'}}
   return SVE_ACLE_FUNC(svqshrunt,_n_s64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svqshrunt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svqshrunt_n_s64_1u12__SVUint32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqshrunt_n_s64_1(svuint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqshrunt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqshrunt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqshrunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqshrunt_n_s64'}}
   return SVE_ACLE_FUNC(svqshrunt,_n_s64,,)(op, op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsub.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsub.c
index 2522cc4aacf71..f669fd9d2cbaa 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsub.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsub.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqsub_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsub_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s8_z'}}
   return SVE_ACLE_FUNC(svqsub,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqsub_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s16_z'}}
   return SVE_ACLE_FUNC(svqsub,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqsub_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s32_z'}}
   return SVE_ACLE_FUNC(svqsub,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqsub_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s64_z'}}
   return SVE_ACLE_FUNC(svqsub,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsub_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u8_z'}}
   return SVE_ACLE_FUNC(svqsub,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqsub_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u16_z'}}
   return SVE_ACLE_FUNC(svqsub,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqsub_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u32_z'}}
   return SVE_ACLE_FUNC(svqsub,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqsub_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u64_z'}}
   return SVE_ACLE_FUNC(svqsub,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqsub_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s8_m'}}
   return SVE_ACLE_FUNC(svqsub,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqsub_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s16_m'}}
   return SVE_ACLE_FUNC(svqsub,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqsub_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s32_m'}}
   return SVE_ACLE_FUNC(svqsub,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqsub_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s64_m'}}
   return SVE_ACLE_FUNC(svqsub,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqsub_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u8_m'}}
   return SVE_ACLE_FUNC(svqsub,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqsub_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u16_m'}}
   return SVE_ACLE_FUNC(svqsub,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqsub_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u32_m'}}
   return SVE_ACLE_FUNC(svqsub,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqsub_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u64_m'}}
   return SVE_ACLE_FUNC(svqsub,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqsub_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s8_x'}}
   return SVE_ACLE_FUNC(svqsub,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqsub_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s16_x'}}
   return SVE_ACLE_FUNC(svqsub,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqsub_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s32_x'}}
   return SVE_ACLE_FUNC(svqsub,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqsub_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_s64_x'}}
   return SVE_ACLE_FUNC(svqsub,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqsub_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqsub_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u8_x'}}
   return SVE_ACLE_FUNC(svqsub,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqsub_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u16_x'}}
   return SVE_ACLE_FUNC(svqsub,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqsub_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u32_x'}}
   return SVE_ACLE_FUNC(svqsub,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsub_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqsub_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_u64_x'}}
   return SVE_ACLE_FUNC(svqsub,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsub_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svqsub_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s8_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svqsub_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s16_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svqsub_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s32_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svqsub_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s64_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsub_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svqsub_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u8_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svqsub_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u16_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svqsub_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u32_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svqsub_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u64_z'}}
   return SVE_ACLE_FUNC(svqsub,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsub_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsub_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s8_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqsub_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s16_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqsub_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s32_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqsub_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s64_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsub_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsub_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u8_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqsub_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u16_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqsub_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u32_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqsub_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u64_m'}}
   return SVE_ACLE_FUNC(svqsub,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsub_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsub_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s8_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqsub_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s16_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqsub_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s32_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqsub_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_s64_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsub_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsub_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u8_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqsub_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u16_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqsub_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u32_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsub_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsub_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqsub_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsub_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsub_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsub_n_u64_x'}}
   return SVE_ACLE_FUNC(svqsub,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsubr.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsubr.c
index 2150fd13f69c5..69784968550f6 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsubr.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsubr.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqsubr_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsubr_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsubr_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s8_z'}}
   return SVE_ACLE_FUNC(svqsubr,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqsubr_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s16_z'}}
   return SVE_ACLE_FUNC(svqsubr,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqsubr_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s32_z'}}
   return SVE_ACLE_FUNC(svqsubr,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqsubr_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s64_z'}}
   return SVE_ACLE_FUNC(svqsubr,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsubr_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsubr_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u8_z'}}
   return SVE_ACLE_FUNC(svqsubr,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqsubr_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u16_z'}}
   return SVE_ACLE_FUNC(svqsubr,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqsubr_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u32_z'}}
   return SVE_ACLE_FUNC(svqsubr,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqsubr_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u64_z'}}
   return SVE_ACLE_FUNC(svqsubr,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsubr_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqsubr_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s8_m'}}
   return SVE_ACLE_FUNC(svqsubr,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqsubr_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s16_m'}}
   return SVE_ACLE_FUNC(svqsubr,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqsubr_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s32_m'}}
   return SVE_ACLE_FUNC(svqsubr,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqsubr_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s64_m'}}
   return SVE_ACLE_FUNC(svqsubr,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsubr_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqsubr_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u8_m'}}
   return SVE_ACLE_FUNC(svqsubr,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqsubr_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u16_m'}}
   return SVE_ACLE_FUNC(svqsubr,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqsubr_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u32_m'}}
   return SVE_ACLE_FUNC(svqsubr,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqsubr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u64_m'}}
   return SVE_ACLE_FUNC(svqsubr,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsubr_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqsubr_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s8_x'}}
   return SVE_ACLE_FUNC(svqsubr,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svqsubr_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s16_x'}}
   return SVE_ACLE_FUNC(svqsubr,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svqsubr_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s32_x'}}
   return SVE_ACLE_FUNC(svqsubr,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svqsubr_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_s64_x'}}
   return SVE_ACLE_FUNC(svqsubr,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqsubr_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqsubr_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u8_x'}}
   return SVE_ACLE_FUNC(svqsubr,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svqsubr_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u16_x'}}
   return SVE_ACLE_FUNC(svqsubr,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svqsubr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u32_x'}}
   return SVE_ACLE_FUNC(svqsubr,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svqsubr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svqsubr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_u64_x'}}
   return SVE_ACLE_FUNC(svqsubr,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsubr_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svqsubr_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s8_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svqsubr_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s16_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svqsubr_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s32_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svqsubr_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s64_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsubr_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svqsubr_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u8_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svqsubr_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u16_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u16_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svqsubr_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u32_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u32_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svqsubr_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u64_z
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u64_z'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsubr_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsubr_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s8_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqsubr_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s16_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqsubr_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s32_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqsubr_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s64_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsubr_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsubr_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u8_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqsubr_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u16_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u16_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqsubr_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u32_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u32_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqsubr_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u64_m
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u64_m'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsubr_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svqsubr_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s8_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svqsubr_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s16_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svqsubr_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s32_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svqsubr_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_s64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_s64_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svqsubr_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svqsubr_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u8_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svqsubr_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u16_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u16_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svqsubr_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u32_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u32_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svqsubr_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svqsubr_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svqsubr_n_u64_x
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqsubr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svqsubr_n_u64_x'}}
   return SVE_ACLE_FUNC(svqsubr,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnb.c
index 267a953c2c3cd..bc6047fd65481 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqxtnb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnb_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqxtnb_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqxtnb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnb.nxv8i16(<vscale x 8 x i16> %op1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnb_s16'}}
   return SVE_ACLE_FUNC(svqxtnb,_s16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtnb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnb_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqxtnb_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqxtnb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnb.nxv4i32(<vscale x 4 x i32> %op1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnb_s32'}}
   return SVE_ACLE_FUNC(svqxtnb,_s32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtnb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnb_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqxtnb_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqxtnb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnb.nxv2i64(<vscale x 2 x i64> %op1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnb_s64'}}
   return SVE_ACLE_FUNC(svqxtnb,_s64,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtnb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnb_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqxtnb_u16(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqxtnb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnb.nxv8i16(<vscale x 8 x i16> %op1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnb_u16'}}
   return SVE_ACLE_FUNC(svqxtnb,_u16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtnb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnb_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqxtnb_u32(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqxtnb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnb.nxv4i32(<vscale x 4 x i32> %op1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnb_u32'}}
   return SVE_ACLE_FUNC(svqxtnb,_u32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtnb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnb_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqxtnb_u64(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqxtnb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnb.nxv2i64(<vscale x 2 x i64> %op1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnb_u64'}}
   return SVE_ACLE_FUNC(svqxtnb,_u64,,)(op1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnt.c
index 5c1da4ba561e7..0cfd9d3cdf631 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,61 +15,103 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqxtnt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnt_s16u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svqxtnt_s16(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqxtnt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnt_s16'}}
   return SVE_ACLE_FUNC(svqxtnt,_s16,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtnt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnt_s32u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svqxtnt_s32(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqxtnt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnt_s32'}}
   return SVE_ACLE_FUNC(svqxtnt,_s32,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtnt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnt_s64u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svqxtnt_s64(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqxtnt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnt_s64'}}
   return SVE_ACLE_FUNC(svqxtnt,_s64,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtnt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnt_u16u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqxtnt_u16(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svqxtnt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnt_u16'}}
   return SVE_ACLE_FUNC(svqxtnt,_u16,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtnt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnt_u32u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqxtnt_u32(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svqxtnt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnt_u32'}}
   return SVE_ACLE_FUNC(svqxtnt,_u32,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtnt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svqxtnt_u64u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqxtnt_u64(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svqxtnt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtnt_u64'}}
   return SVE_ACLE_FUNC(svqxtnt,_u64,,)(op, op1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunb.c
index 7fa498ee212bd..cb244a4f98843 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,31 +15,52 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqxtunb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqxtunb_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqxtunb_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svqxtunb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunb.nxv8i16(<vscale x 8 x i16> %op1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtunb_s16'}}
   return SVE_ACLE_FUNC(svqxtunb,_s16,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtunb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqxtunb_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqxtunb_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svqxtunb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunb.nxv4i32(<vscale x 4 x i32> %op1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtunb_s32'}}
   return SVE_ACLE_FUNC(svqxtunb,_s32,,)(op1);
 }
 
+// CHECK-LABEL: @test_svqxtunb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqxtunb_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqxtunb_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svqxtunb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunb.nxv2i64(<vscale x 2 x i64> %op1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtunb'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtunb_s64'}}
   return SVE_ACLE_FUNC(svqxtunb,_s64,,)(op1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunt.c
index d039e8c92ddad..b4ea2415ac21e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qxtunt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,31 +15,52 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svqxtunt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqxtunt_u16u11__SVUint8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svqxtunt_u16(svuint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svqxtunt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtunt_s16'}}
   return SVE_ACLE_FUNC(svqxtunt,_s16,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtunt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqxtunt_u32u12__SVUint16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svqxtunt_u32(svuint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svqxtunt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtunt_s32'}}
   return SVE_ACLE_FUNC(svqxtunt,_s32,,)(op, op1);
 }
 
+// CHECK-LABEL: @test_svqxtunt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svqxtunt_u64u12__SVUint32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svqxtunt_u64(svuint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svqxtunt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svqxtunt'}}
   // expected-warning at +1 {{implicit declaration of function 'svqxtunt_s64'}}
   return SVE_ACLE_FUNC(svqxtunt,_s64,,)(op, op1);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnb.c
index 204731fce79ef..af12457ee4799 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svraddhnb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnb_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svraddhnb_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_s16'}}
   return SVE_ACLE_FUNC(svraddhnb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnb_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svraddhnb_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_s32'}}
   return SVE_ACLE_FUNC(svraddhnb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnb_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svraddhnb_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_s64'}}
   return SVE_ACLE_FUNC(svraddhnb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnb_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svraddhnb_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_u16'}}
   return SVE_ACLE_FUNC(svraddhnb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnb_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svraddhnb_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_u32'}}
   return SVE_ACLE_FUNC(svraddhnb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnb_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svraddhnb_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_u64'}}
   return SVE_ACLE_FUNC(svraddhnb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnb_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svraddhnb_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_n_s16'}}
   return SVE_ACLE_FUNC(svraddhnb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnb_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svraddhnb_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_n_s32'}}
   return SVE_ACLE_FUNC(svraddhnb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnb_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svraddhnb_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_n_s64'}}
   return SVE_ACLE_FUNC(svraddhnb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnb_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svraddhnb_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_n_u16'}}
   return SVE_ACLE_FUNC(svraddhnb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnb_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svraddhnb_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_n_u32'}}
   return SVE_ACLE_FUNC(svraddhnb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svraddhnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnb_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svraddhnb_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svraddhnb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnb_n_u64'}}
   return SVE_ACLE_FUNC(svraddhnb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnt.c
index 15c0e9025b508..8d42071762085 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_raddhnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svraddhnt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnt_s16u10__SVInt8_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svraddhnt_s16(svint8_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_s16'}}
   return SVE_ACLE_FUNC(svraddhnt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnt_s32u11__SVInt16_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svraddhnt_s32(svint16_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_s32'}}
   return SVE_ACLE_FUNC(svraddhnt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnt_s64u11__SVInt32_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svraddhnt_s64(svint32_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_s64'}}
   return SVE_ACLE_FUNC(svraddhnt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnt_u16u11__SVUint8_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svraddhnt_u16(svuint8_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_u16'}}
   return SVE_ACLE_FUNC(svraddhnt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnt_u32u12__SVUint16_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svraddhnt_u32(svuint16_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_u32'}}
   return SVE_ACLE_FUNC(svraddhnt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svraddhnt_u64u12__SVUint32_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svraddhnt_u64(svuint32_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_u64'}}
   return SVE_ACLE_FUNC(svraddhnt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnt_n_s16u10__SVInt8_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svraddhnt_n_s16(svint8_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_n_s16'}}
   return SVE_ACLE_FUNC(svraddhnt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnt_n_s32u11__SVInt16_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svraddhnt_n_s32(svint16_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_n_s32'}}
   return SVE_ACLE_FUNC(svraddhnt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnt_n_s64u11__SVInt32_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svraddhnt_n_s64(svint32_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_n_s64'}}
   return SVE_ACLE_FUNC(svraddhnt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnt_n_u16u11__SVUint8_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svraddhnt_n_u16(svuint8_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.raddhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_n_u16'}}
   return SVE_ACLE_FUNC(svraddhnt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnt_n_u32u12__SVUint16_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svraddhnt_n_u32(svuint16_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.raddhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_n_u32'}}
   return SVE_ACLE_FUNC(svraddhnt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svraddhnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svraddhnt_n_u64u12__SVUint32_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svraddhnt_n_u64(svuint32_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svraddhnt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.raddhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svraddhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svraddhnt_n_u64'}}
   return SVE_ACLE_FUNC(svraddhnt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rax1.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rax1.c
index 7fdb0bc3ba6f8..41d878086c14e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rax1.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rax1.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sha3 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sha3 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sha3 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sha3 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sha3 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sha3 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,21 +15,35 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrax1_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rax1(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svrax1_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rax1(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svrax1_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrax1_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rax1(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrax1'}}
   // expected-warning at +1 {{implicit declaration of function 'svrax1_s64'}}
   return SVE_ACLE_FUNC(svrax1,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrax1_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rax1(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svrax1_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rax1(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svrax1_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrax1_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.rax1(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrax1'}}
   // expected-warning at +1 {{implicit declaration of function 'svrax1_u64'}}
   return SVE_ACLE_FUNC(svrax1,_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_recpe.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_recpe.c
index 19a88273caef6..b3858698e02c2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_recpe.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_recpe.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,34 +15,58 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrecpe_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpe_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrecpe_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrecpe_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrecpe_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrecpe_u32_z'}}
   return SVE_ACLE_FUNC(svrecpe,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpe_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpe_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrecpe_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrecpe_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrecpe_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrecpe_u32_m'}}
   return SVE_ACLE_FUNC(svrecpe,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrecpe_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrecpe_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrecpe_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrecpe_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urecpe.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrecpe_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrecpe_u32_x'}}
   return SVE_ACLE_FUNC(svrecpe,_u32,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rhadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rhadd.c
index 6e4f5ca98aa80..d37487073a7e8 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rhadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rhadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 #include <arm_sve.h>
@@ -13,557 +14,972 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrhadd_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrhadd_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrhadd_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s8_m'}}
   return SVE_ACLE_FUNC(svrhadd,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrhadd_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s16_m'}}
   return SVE_ACLE_FUNC(svrhadd,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrhadd_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s32_m'}}
   return SVE_ACLE_FUNC(svrhadd,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrhadd_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s64_m'}}
   return SVE_ACLE_FUNC(svrhadd,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrhadd_u8_mu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrhadd_u8_m(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u8_m'}}
   return SVE_ACLE_FUNC(svrhadd,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u16_mu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrhadd_u16_m(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u16_m'}}
   return SVE_ACLE_FUNC(svrhadd,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u32_mu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrhadd_u32_m(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svrhadd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u32_m'}}
   return SVE_ACLE_FUNC(svrhadd,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u64_mu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrhadd_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u64_m'}}
   return SVE_ACLE_FUNC(svrhadd,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrhadd_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrhadd_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s8_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrhadd_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s16_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrhadd_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s32_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrhadd_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s64_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrhadd_n_u8_mu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrhadd_n_u8_m(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u8_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u16_mu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrhadd_n_u16_m(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u16_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u32_mu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrhadd_n_u32_m(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u32_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u64_mu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrhadd_n_u64_m(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u64_m'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrhadd_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrhadd_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s8_z'}}
   return SVE_ACLE_FUNC(svrhadd,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrhadd_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s16_z'}}
   return SVE_ACLE_FUNC(svrhadd,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrhadd_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s32_z'}}
   return SVE_ACLE_FUNC(svrhadd,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrhadd_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s64_z'}}
   return SVE_ACLE_FUNC(svrhadd,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrhadd_u8_zu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrhadd_u8_z(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u8_z'}}
   return SVE_ACLE_FUNC(svrhadd,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u16_zu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrhadd_u16_z(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u16_z'}}
   return SVE_ACLE_FUNC(svrhadd,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u32_zu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrhadd_u32_z(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svrhadd_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u32_z'}}
   return SVE_ACLE_FUNC(svrhadd,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u64_zu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrhadd_u64_z(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u64_z'}}
   return SVE_ACLE_FUNC(svrhadd,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrhadd_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svrhadd_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s8_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svrhadd_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s16_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svrhadd_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s32_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svrhadd_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s64_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrhadd_n_u8_zu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svrhadd_n_u8_z(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u8_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u16_zu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svrhadd_n_u16_z(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u16_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u32_zu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svrhadd_n_u32_z(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u32_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u64_zu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svrhadd_n_u64_z(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u64_z'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrhadd_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrhadd_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s8_x'}}
   return SVE_ACLE_FUNC(svrhadd,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrhadd_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s16_x'}}
   return SVE_ACLE_FUNC(svrhadd,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrhadd_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s32_x'}}
   return SVE_ACLE_FUNC(svrhadd,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrhadd_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_s64_x'}}
   return SVE_ACLE_FUNC(svrhadd,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrhadd_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrhadd_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u8_x'}}
   return SVE_ACLE_FUNC(svrhadd,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrhadd_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u16_x'}}
   return SVE_ACLE_FUNC(svrhadd,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrhadd_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svrhadd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u32_x'}}
   return SVE_ACLE_FUNC(svrhadd,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrhadd_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrhadd_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_u64_x'}}
   return SVE_ACLE_FUNC(svrhadd,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrhadd_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrhadd_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s8_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrhadd_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s16_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrhadd_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s32_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrhadd_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_s64_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrhadd_n_u8_xu10__SVBool_tu11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrhadd_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urhadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u8_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u16_xu10__SVBool_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrhadd_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urhadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u16_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u32_xu10__SVBool_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrhadd_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urhadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u32_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrhadd_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrhadd_n_u64_xu10__SVBool_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrhadd_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svrhadd_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urhadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrhadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrhadd_n_u64_x'}}
   return SVE_ACLE_FUNC(svrhadd,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshl.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshl.c
index 4debc2d6555e1..6df89c4aace80 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshl.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshl.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,557 +15,969 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrshl_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrshl_s8_zu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrshl_s8_z(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s8_z'}}
   return SVE_ACLE_FUNC(svrshl,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s16_zu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrshl_s16_z(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s16_z'}}
   return SVE_ACLE_FUNC(svrshl,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s32_zu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrshl_s32_z(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s32_z'}}
   return SVE_ACLE_FUNC(svrshl,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s64_zu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrshl_s64_z(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s64_z'}}
   return SVE_ACLE_FUNC(svrshl,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrshl_u8_zu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrshl_u8_z(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u8_z'}}
   return SVE_ACLE_FUNC(svrshl,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u16_zu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrshl_u16_z(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u16_z'}}
   return SVE_ACLE_FUNC(svrshl,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u32_zu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrshl_u32_z(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u32_z'}}
   return SVE_ACLE_FUNC(svrshl,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u64_zu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrshl_u64_z(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u64_z'}}
   return SVE_ACLE_FUNC(svrshl,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrshl_s8_mu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshl_s8_m(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s8_m'}}
   return SVE_ACLE_FUNC(svrshl,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s16_mu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrshl_s16_m(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s16_m'}}
   return SVE_ACLE_FUNC(svrshl,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s32_mu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrshl_s32_m(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s32_m'}}
   return SVE_ACLE_FUNC(svrshl,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s64_mu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrshl_s64_m(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s64_m'}}
   return SVE_ACLE_FUNC(svrshl,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrshl_u8_mu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshl_u8_m(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u8_m'}}
   return SVE_ACLE_FUNC(svrshl,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u16_mu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrshl_u16_m(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u16_m'}}
   return SVE_ACLE_FUNC(svrshl,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u32_mu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrshl_u32_m(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u32_m'}}
   return SVE_ACLE_FUNC(svrshl,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u64_mu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrshl_u64_m(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u64_m'}}
   return SVE_ACLE_FUNC(svrshl,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrshl_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshl_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s8_x'}}
   return SVE_ACLE_FUNC(svrshl,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrshl_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s16_x'}}
   return SVE_ACLE_FUNC(svrshl,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrshl_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s32_x'}}
   return SVE_ACLE_FUNC(svrshl,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrshl_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_s64_x'}}
   return SVE_ACLE_FUNC(svrshl,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrshl_u8_xu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshl_u8_x(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u8_x'}}
   return SVE_ACLE_FUNC(svrshl,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u16_xu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrshl_u16_x(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u16_x'}}
   return SVE_ACLE_FUNC(svrshl,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u32_xu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrshl_u32_x(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u32_x'}}
   return SVE_ACLE_FUNC(svrshl,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrshl_u64_xu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrshl_u64_x(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_u64_x'}}
   return SVE_ACLE_FUNC(svrshl,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshl_n_s8_zu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svrshl_n_s8_z(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s8_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s16_zu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svrshl_n_s16_z(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s16_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s32_zu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svrshl_n_s32_z(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s32_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s64_zu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svrshl_n_s64_z(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s64_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshl_n_u8_zu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svrshl_n_u8_z(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u8_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u8_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u16_zu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svrshl_n_u16_z(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u16_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u32_zu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svrshl_n_u32_z(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u32_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u64_zu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svrshl_n_u64_z(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u64_z'}}
   return SVE_ACLE_FUNC(svrshl,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshl_n_s8_mu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrshl_n_s8_m(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s8_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s16_mu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrshl_n_s16_m(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s16_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s32_mu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrshl_n_s32_m(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s32_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s64_mu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrshl_n_s64_m(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s64_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshl_n_u8_mu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrshl_n_u8_m(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u8_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u16_mu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrshl_n_u16_m(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u16_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u32_mu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrshl_n_u32_m(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u32_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u64_mu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrshl_n_u64_m(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u64_m'}}
   return SVE_ACLE_FUNC(svrshl,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshl_n_s8_xu10__SVBool_tu10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrshl_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s8_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s16_xu10__SVBool_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrshl_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s16_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s32_xu10__SVBool_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrshl_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s32_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_s64_xu10__SVBool_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrshl_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_s64_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshl_n_u8_xu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrshl_n_u8_x(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshl.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u8_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u16_xu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrshl_n_u16_x(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshl.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u16_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u32_xu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrshl_n_u32_x(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshl.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u32_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svrshl_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshl_n_u64_xu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrshl_n_u64_x(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrshl_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshl.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshl_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshl_n_u64_x'}}
   return SVE_ACLE_FUNC(svrshl,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshr.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshr.c
index 8923be4528485..0ffac50e37f06 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshr.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshr.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,533 +15,921 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrshr_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshr_n_s8_zu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrshr_n_s8_z(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s8_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s8,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s8_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrshr_n_s8_z_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrshr_n_s8_z_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s8_z_1
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s8_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s8,_z,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s16_zu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrshr_n_s16_z(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s16_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s16,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s16_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s16_z_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svrshr_n_s16_z_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s16_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s16_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s16,_z,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s32_zu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrshr_n_s32_z(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s32_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s32,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s32_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s32_z_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svrshr_n_s32_z_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s32_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s32_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s32,_z,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s64_zu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrshr_n_s64_z(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s64_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s64,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s64_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s64_z_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svrshr_n_s64_z_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s64_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s64_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_s64,_z,)(pg, op1, 64);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshr_n_u8_zu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrshr_n_u8_z(svbool_t pg, svuint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u8_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u8,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u8_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrshr_n_u8_z_1u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrshr_n_u8_z_1(svbool_t pg, svuint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u8_z_1
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u8_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u8,_z,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u16_zu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrshr_n_u16_z(svbool_t pg, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u16_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u16,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u16_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u16_z_1u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svrshr_n_u16_z_1(svbool_t pg, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u16_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u16_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u16,_z,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrshr_n_u32_z(svbool_t pg, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u32_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u32,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u32_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u32_z_1u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svrshr_n_u32_z_1(svbool_t pg, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u32_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u32_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u32,_z,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u64_zu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrshr_n_u64_z(svbool_t pg, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u64_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u64,_z,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u64_z_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u64_z_1u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svrshr_n_u64_z_1(svbool_t pg, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u64_z_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u64_z'}}
   return SVE_ACLE_FUNC(svrshr,_n_u64,_z,)(pg, op1, 64);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshr_n_s8_mu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshr_n_s8_m(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s8_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s8,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s8_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrshr_n_s8_m_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshr_n_s8_m_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s8_m_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s8_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s8,_m,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s16_mu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrshr_n_s16_m(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s16_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s16,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s16_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s16_m_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrshr_n_s16_m_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s16_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s16_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s16,_m,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s32_mu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrshr_n_s32_m(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s32_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s32,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s32_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s32_m_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrshr_n_s32_m_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s32_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s32_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s32,_m,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s64_mu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrshr_n_s64_m(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s64_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s64,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s64_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s64_m_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrshr_n_s64_m_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s64_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s64_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_s64,_m,)(pg, op1, 64);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshr_n_u8_mu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshr_n_u8_m(svbool_t pg, svuint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u8_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u8,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u8_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrshr_n_u8_m_1u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshr_n_u8_m_1(svbool_t pg, svuint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u8_m_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u8_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u8,_m,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u16_mu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrshr_n_u16_m(svbool_t pg, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u16_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u16,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u16_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u16_m_1u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrshr_n_u16_m_1(svbool_t pg, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u16_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u16_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u16,_m,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u32_mu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrshr_n_u32_m(svbool_t pg, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u32_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u32,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u32_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u32_m_1u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrshr_n_u32_m_1(svbool_t pg, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u32_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u32_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u32,_m,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u64_mu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrshr_n_u64_m(svbool_t pg, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u64_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u64,_m,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u64_m_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u64_m_1u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrshr_n_u64_m_1(svbool_t pg, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u64_m_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u64_m'}}
   return SVE_ACLE_FUNC(svrshr,_n_u64,_m,)(pg, op1, 64);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshr_n_s8_xu10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshr_n_s8_x(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s8_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s8,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s8_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrshr_n_s8_x_1u10__SVBool_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshr_n_s8_x_1(svbool_t pg, svint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s8_x_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s8_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s8,_x,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s16_xu10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrshr_n_s16_x(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s16_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s16,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s16_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s16_x_1u10__SVBool_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrshr_n_s16_x_1(svbool_t pg, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s16_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s16_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s16,_x,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s32_xu10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrshr_n_s32_x(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s32_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s32,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s32_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s32_x_1u10__SVBool_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrshr_n_s32_x_1(svbool_t pg, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s32_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s32_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s32,_x,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_s64_xu10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrshr_n_s64_x(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s64_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s64,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_s64_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_s64_x_1u10__SVBool_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svrshr_n_s64_x_1(svbool_t pg, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_s64_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_s64_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_s64,_x,)(pg, op1, 64);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrshr_n_u8_xu10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshr_n_u8_x(svbool_t pg, svuint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u8_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u8,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u8_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrshr_n_u8_x_1u10__SVBool_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshr_n_u8_x_1(svbool_t pg, svuint8_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u8_x_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.urshr.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u8_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u8,_x,)(pg, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u16_xu10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrshr_n_u16_x(svbool_t pg, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u16_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u16,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u16_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u16_x_1u10__SVBool_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrshr_n_u16_x_1(svbool_t pg, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u16_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.urshr.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u16_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u16,_x,)(pg, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrshr_n_u32_x(svbool_t pg, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u32_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u32,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u32_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u32_x_1u10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrshr_n_u32_x_1(svbool_t pg, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u32_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.urshr.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u32_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u32,_x,)(pg, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshr_n_u64_xu10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrshr_n_u64_x(svbool_t pg, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u64_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u64,_x,)(pg, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshr_n_u64_x_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshr_n_u64_x_1u10__SVBool_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svrshr_n_u64_x_1(svbool_t pg, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshr_n_u64_x_1
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.urshr.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshr_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshr_n_u64_x'}}
   return SVE_ACLE_FUNC(svrshr,_n_u64,_x,)(pg, op1, 64);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnb.c
index 031c1381536df..0c227a431c75f 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrshrnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnb_n_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshrnb_n_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_s16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnb_n_s16_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshrnb_n_s16_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_s16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnb_n_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrshrnb_n_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_s32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnb_n_s32_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrshrnb_n_s32_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_s32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnb_n_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrshrnb_n_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_s64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnb_n_s64_1u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrshrnb_n_s64_1(svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_s64,,)(op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnb_n_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshrnb_n_u16(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_u16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnb_n_u16_1u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshrnb_n_u16_1(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_u16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnb_n_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrshrnb_n_u32(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_u32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnb_n_u32_1u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrshrnb_n_u32_1(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_u32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnb_n_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrshrnb_n_u64(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_u64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnb_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnb_n_u64_1u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrshrnb_n_u64_1(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnb_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svrshrnb,_n_u64,,)(op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnt.c
index a46ac59b7c4b7..fbc26c2b27af6 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rshrnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrshrnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnt_n_s16u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshrnt_n_s16(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_s16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnt_n_s16_1u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrshrnt_n_s16_1(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_s16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnt_n_s32u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrshrnt_n_s32(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_s32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnt_n_s32_1u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrshrnt_n_s32_1(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_s32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnt_n_s64u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrshrnt_n_s64(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_s64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnt_n_s64_1u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrshrnt_n_s64_1(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_s64,,)(op, op1, 32);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnt_n_u16u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshrnt_n_u16(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_u16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnt_n_u16_1u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrshrnt_n_u16_1(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rshrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_u16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnt_n_u32u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrshrnt_n_u32(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_u32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnt_n_u32_1u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrshrnt_n_u32_1(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rshrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_u32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrshrnt_n_u64u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrshrnt_n_u64(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_u64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svrshrnt_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svrshrnt_n_u64_1u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrshrnt_n_u64_1(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svrshrnt_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rshrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svrshrnt,_n_u64,,)(op, op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsqrte.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsqrte.c
index 528c838d1fc5e..9cc80cad7af14 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsqrte.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsqrte.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,34 +15,58 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrsqrte_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsqrte_u32_zu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrsqrte_u32_z(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrsqrte_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> zeroinitializer, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsqrte_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsqrte_u32_z'}}
   return SVE_ACLE_FUNC(svrsqrte,_u32,_z,)(pg, op);
 }
 
+// CHECK-LABEL: @test_svrsqrte_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsqrte_u32_mu12__SVUint32_tu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> [[INACTIVE:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrsqrte_u32_m(svuint32_t inactive, svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrsqrte_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> %inactive, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsqrte_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsqrte_u32_m'}}
   return SVE_ACLE_FUNC(svrsqrte,_u32,_m,)(inactive, pg, op);
 }
 
+// CHECK-LABEL: @test_svrsqrte_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsqrte_u32_xu10__SVBool_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrsqrte_u32_x(svbool_t pg, svuint32_t op)
 {
-  // CHECK-LABEL: test_svrsqrte_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursqrte.nxv4i32(<vscale x 4 x i32> undef, <vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsqrte_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsqrte_u32_x'}}
   return SVE_ACLE_FUNC(svrsqrte,_u32,_x,)(pg, op);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsra.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsra.c
index 50af90afc5a90..cc76cc6d7d6ed 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsra.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsra.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,161 +15,273 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrsra_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srsra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrsra_n_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srsra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrsra_n_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srsra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s8'}}
   return SVE_ACLE_FUNC(svrsra,_n_s8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srsra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsra_n_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srsra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrsra_n_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.srsra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s8'}}
   return SVE_ACLE_FUNC(svrsra,_n_s8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srsra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsra_n_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srsra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrsra_n_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srsra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s16'}}
   return SVE_ACLE_FUNC(svrsra,_n_s16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srsra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsra_n_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srsra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrsra_n_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.srsra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s16'}}
   return SVE_ACLE_FUNC(svrsra,_n_s16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srsra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsra_n_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srsra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrsra_n_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srsra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s32'}}
   return SVE_ACLE_FUNC(svrsra,_n_s32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srsra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsra_n_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srsra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrsra_n_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.srsra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s32'}}
   return SVE_ACLE_FUNC(svrsra,_n_s32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srsra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsra_n_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srsra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svrsra_n_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srsra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s64'}}
   return SVE_ACLE_FUNC(svrsra,_n_s64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srsra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsra_n_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srsra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svrsra_n_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.srsra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_s64'}}
   return SVE_ACLE_FUNC(svrsra,_n_s64,,)(op1, op2, 64);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ursra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svrsra_n_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ursra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrsra_n_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ursra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u8'}}
   return SVE_ACLE_FUNC(svrsra,_n_u8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ursra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsra_n_u8_1u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ursra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrsra_n_u8_1(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ursra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u8'}}
   return SVE_ACLE_FUNC(svrsra,_n_u8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ursra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsra_n_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ursra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrsra_n_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ursra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u16'}}
   return SVE_ACLE_FUNC(svrsra,_n_u16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ursra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsra_n_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ursra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrsra_n_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ursra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u16'}}
   return SVE_ACLE_FUNC(svrsra,_n_u16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsra_n_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrsra_n_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u32'}}
   return SVE_ACLE_FUNC(svrsra,_n_u32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsra_n_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrsra_n_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ursra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u32'}}
   return SVE_ACLE_FUNC(svrsra,_n_u32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ursra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svrsra_n_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ursra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svrsra_n_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ursra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u64'}}
   return SVE_ACLE_FUNC(svrsra,_n_u64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svrsra_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ursra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svrsra_n_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ursra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svrsra_n_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrsra_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ursra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsra_n_u64'}}
   return SVE_ACLE_FUNC(svrsra,_n_u64,,)(op1, op2, 64);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnb.c
index 8e7532c66f262..10f81f0e65538 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrsubhnb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnb_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrsubhnb_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_s16'}}
   return SVE_ACLE_FUNC(svrsubhnb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnb_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrsubhnb_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_s32'}}
   return SVE_ACLE_FUNC(svrsubhnb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnb_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrsubhnb_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_s64'}}
   return SVE_ACLE_FUNC(svrsubhnb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnb_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrsubhnb_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_u16'}}
   return SVE_ACLE_FUNC(svrsubhnb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnb_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrsubhnb_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_u32'}}
   return SVE_ACLE_FUNC(svrsubhnb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnb_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrsubhnb_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_u64'}}
   return SVE_ACLE_FUNC(svrsubhnb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnb_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrsubhnb_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_n_s16'}}
   return SVE_ACLE_FUNC(svrsubhnb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnb_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrsubhnb_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_n_s32'}}
   return SVE_ACLE_FUNC(svrsubhnb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnb_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrsubhnb_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_n_s64'}}
   return SVE_ACLE_FUNC(svrsubhnb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnb_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrsubhnb_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_n_u16'}}
   return SVE_ACLE_FUNC(svrsubhnb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnb_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrsubhnb_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_n_u32'}}
   return SVE_ACLE_FUNC(svrsubhnb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svrsubhnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnb_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrsubhnb_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svrsubhnb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnb_n_u64'}}
   return SVE_ACLE_FUNC(svrsubhnb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnt.c
index 0ca5ca1ea1ef4..27084aaba0d80 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_rsubhnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svrsubhnt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnt_s16u10__SVInt8_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svrsubhnt_s16(svint8_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_s16'}}
   return SVE_ACLE_FUNC(svrsubhnt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnt_s32u11__SVInt16_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svrsubhnt_s32(svint16_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_s32'}}
   return SVE_ACLE_FUNC(svrsubhnt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnt_s64u11__SVInt32_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svrsubhnt_s64(svint32_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_s64'}}
   return SVE_ACLE_FUNC(svrsubhnt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnt_u16u11__SVUint8_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svrsubhnt_u16(svuint8_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_u16'}}
   return SVE_ACLE_FUNC(svrsubhnt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnt_u32u12__SVUint16_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svrsubhnt_u32(svuint16_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_u32'}}
   return SVE_ACLE_FUNC(svrsubhnt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svrsubhnt_u64u12__SVUint32_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svrsubhnt_u64(svuint32_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_u64'}}
   return SVE_ACLE_FUNC(svrsubhnt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnt_n_s16u10__SVInt8_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svrsubhnt_n_s16(svint8_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_n_s16'}}
   return SVE_ACLE_FUNC(svrsubhnt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnt_n_s32u11__SVInt16_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svrsubhnt_n_s32(svint16_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_n_s32'}}
   return SVE_ACLE_FUNC(svrsubhnt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnt_n_s64u11__SVInt32_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svrsubhnt_n_s64(svint32_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_n_s64'}}
   return SVE_ACLE_FUNC(svrsubhnt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnt_n_u16u11__SVUint8_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svrsubhnt_n_u16(svuint8_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.rsubhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_n_u16'}}
   return SVE_ACLE_FUNC(svrsubhnt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnt_n_u32u12__SVUint16_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svrsubhnt_n_u32(svuint16_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.rsubhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_n_u32'}}
   return SVE_ACLE_FUNC(svrsubhnt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svrsubhnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svrsubhnt_n_u64u12__SVUint32_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svrsubhnt_n_u64(svuint32_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svrsubhnt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.rsubhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svrsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svrsubhnt_n_u64'}}
   return SVE_ACLE_FUNC(svrsubhnt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclb.c
index 7686f3d6f1edd..821bb94873fe2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,43 +15,73 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsbclb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsbclb_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsbclb_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svsbclb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclb_u32'}}
   return SVE_ACLE_FUNC(svsbclb,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsbclb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsbclb_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsbclb_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svsbclb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclb_u64'}}
   return SVE_ACLE_FUNC(svsbclb,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsbclb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsbclb_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsbclb_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svsbclb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclb_n_u32'}}
   return SVE_ACLE_FUNC(svsbclb,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsbclb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsbclb_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsbclb_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svsbclb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclb_n_u64'}}
   return SVE_ACLE_FUNC(svsbclb,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclt.c
index a9c1f3a48251c..85d62531288c6 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sbclt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,43 +15,73 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsbclt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsbclt_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsbclt_u32(svuint32_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svsbclt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclt_u32'}}
   return SVE_ACLE_FUNC(svsbclt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsbclt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsbclt_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsbclt_u64(svuint64_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svsbclt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclt_u64'}}
   return SVE_ACLE_FUNC(svsbclt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsbclt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsbclt_n_u32u12__SVUint32_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsbclt_n_u32(svuint32_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svsbclt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sbclt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclt_n_u32'}}
   return SVE_ACLE_FUNC(svsbclt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsbclt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsbclt_n_u64u12__SVUint64_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsbclt_n_u64(svuint64_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svsbclt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sbclt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsbclt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsbclt_n_u64'}}
   return SVE_ACLE_FUNC(svsbclt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllb.c
index 4d457aab44e11..de856500200b2 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svshllb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllb_n_s16u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshllb_n_s16(svint8_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_s16'}}
   return SVE_ACLE_FUNC(svshllb,_n_s16,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllb_n_s16_1u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshllb_n_s16_1(svint8_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllb.nxv8i16(<vscale x 16 x i8> %op1, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_s16'}}
   return SVE_ACLE_FUNC(svshllb,_n_s16,,)(op1, 7);
 }
 
+// CHECK-LABEL: @test_svshllb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllb_n_s32u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshllb_n_s32(svint16_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_s32'}}
   return SVE_ACLE_FUNC(svshllb,_n_s32,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllb_n_s32_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshllb_n_s32_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllb.nxv4i32(<vscale x 8 x i16> %op1, i32 15)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_s32'}}
   return SVE_ACLE_FUNC(svshllb,_n_s32,,)(op1, 15);
 }
 
+// CHECK-LABEL: @test_svshllb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllb_n_s64u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svshllb_n_s64(svint32_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_s64'}}
   return SVE_ACLE_FUNC(svshllb,_n_s64,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllb_n_s64_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svshllb_n_s64_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllb.nxv2i64(<vscale x 4 x i32> %op1, i32 31)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_s64'}}
   return SVE_ACLE_FUNC(svshllb,_n_s64,,)(op1, 31);
 }
 
+// CHECK-LABEL: @test_svshllb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllb_n_u16u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshllb_n_u16(svuint8_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_u16'}}
   return SVE_ACLE_FUNC(svshllb,_n_u16,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllb_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllb_n_u16_1u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshllb_n_u16_1(svuint8_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllb.nxv8i16(<vscale x 16 x i8> %op1, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_u16'}}
   return SVE_ACLE_FUNC(svshllb,_n_u16,,)(op1, 7);
 }
 
+// CHECK-LABEL: @test_svshllb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllb_n_u32u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshllb_n_u32(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_u32'}}
   return SVE_ACLE_FUNC(svshllb,_n_u32,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllb_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllb_n_u32_1u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshllb_n_u32_1(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllb.nxv4i32(<vscale x 8 x i16> %op1, i32 15)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_u32'}}
   return SVE_ACLE_FUNC(svshllb,_n_u32,,)(op1, 15);
 }
 
+// CHECK-LABEL: @test_svshllb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllb_n_u64u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svshllb_n_u64(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_u64'}}
   return SVE_ACLE_FUNC(svshllb,_n_u64,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllb_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllb_n_u64_1u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svshllb_n_u64_1(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshllb_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllb.nxv2i64(<vscale x 4 x i32> %op1, i32 31)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllb_n_u64'}}
   return SVE_ACLE_FUNC(svshllb,_n_u64,,)(op1, 31);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllt.c
index 2f78ad61545e6..fcf31d60e33bc 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shllt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svshllt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllt_n_s16u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshllt_n_s16(svint8_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_s16'}}
   return SVE_ACLE_FUNC(svshllt,_n_s16,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllt_n_s16_1u10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshllt_n_s16_1(svint8_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sshllt.nxv8i16(<vscale x 16 x i8> %op1, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_s16'}}
   return SVE_ACLE_FUNC(svshllt,_n_s16,,)(op1, 7);
 }
 
+// CHECK-LABEL: @test_svshllt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllt_n_s32u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshllt_n_s32(svint16_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_s32'}}
   return SVE_ACLE_FUNC(svshllt,_n_s32,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllt_n_s32_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshllt_n_s32_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sshllt.nxv4i32(<vscale x 8 x i16> %op1, i32 15)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_s32'}}
   return SVE_ACLE_FUNC(svshllt,_n_s32,,)(op1, 15);
 }
 
+// CHECK-LABEL: @test_svshllt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllt_n_s64u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svshllt_n_s64(svint32_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_s64'}}
   return SVE_ACLE_FUNC(svshllt,_n_s64,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllt_n_s64_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svshllt_n_s64_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sshllt.nxv2i64(<vscale x 4 x i32> %op1, i32 31)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_s64'}}
   return SVE_ACLE_FUNC(svshllt,_n_s64,,)(op1, 31);
 }
 
+// CHECK-LABEL: @test_svshllt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllt_n_u16u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshllt_n_u16(svuint8_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> %op1, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_u16'}}
   return SVE_ACLE_FUNC(svshllt,_n_u16,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllt_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllt_n_u16_1u11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshllt_n_u16_1(svuint8_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ushllt.nxv8i16(<vscale x 16 x i8> %op1, i32 7)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_u16'}}
   return SVE_ACLE_FUNC(svshllt,_n_u16,,)(op1, 7);
 }
 
+// CHECK-LABEL: @test_svshllt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllt_n_u32u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshllt_n_u32(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> %op1, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_u32'}}
   return SVE_ACLE_FUNC(svshllt,_n_u32,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllt_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllt_n_u32_1u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshllt_n_u32_1(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ushllt.nxv4i32(<vscale x 8 x i16> %op1, i32 15)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_u32'}}
   return SVE_ACLE_FUNC(svshllt,_n_u32,,)(op1, 15);
 }
 
+// CHECK-LABEL: @test_svshllt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshllt_n_u64u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svshllt_n_u64(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> %op1, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_u64'}}
   return SVE_ACLE_FUNC(svshllt,_n_u64,,)(op1, 0);
 }
 
+// CHECK-LABEL: @test_svshllt_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshllt_n_u64_1u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svshllt_n_u64_1(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshllt_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ushllt.nxv2i64(<vscale x 4 x i32> %op1, i32 31)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshllt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshllt_n_u64'}}
   return SVE_ACLE_FUNC(svshllt,_n_u64,,)(op1, 31);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnb.c
index 92837b904fc5d..dec3e379bcb9e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svshrnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnb_n_s16u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svshrnb_n_s16(svint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svshrnb,_n_s16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnb_n_s16_1u11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svshrnb_n_s16_1(svint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_s16'}}
   return SVE_ACLE_FUNC(svshrnb,_n_s16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnb_n_s32u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshrnb_n_s32(svint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svshrnb,_n_s32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnb_n_s32_1u11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshrnb_n_s32_1(svint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_s32'}}
   return SVE_ACLE_FUNC(svshrnb,_n_s32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnb_n_s64u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshrnb_n_s64(svint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svshrnb,_n_s64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnb_n_s64_1u11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshrnb_n_s64_1(svint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_s64'}}
   return SVE_ACLE_FUNC(svshrnb,_n_s64,,)(op1, 32);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnb_n_u16u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svshrnb_n_u16(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svshrnb,_n_u16,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnb_n_u16_1u12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svshrnb_n_u16_1(svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnb.nxv8i16(<vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_u16'}}
   return SVE_ACLE_FUNC(svshrnb,_n_u16,,)(op1, 8);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnb_n_u32u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshrnb_n_u32(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svshrnb,_n_u32,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnb_n_u32_1u12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshrnb_n_u32_1(svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnb.nxv4i32(<vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_u32'}}
   return SVE_ACLE_FUNC(svshrnb,_n_u32,,)(op1, 16);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnb_n_u64u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshrnb_n_u64(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svshrnb,_n_u64,,)(op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnb_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnb_n_u64_1u12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshrnb_n_u64_1(svuint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnb_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnb.nxv2i64(<vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnb_n_u64'}}
   return SVE_ACLE_FUNC(svshrnb,_n_u64,,)(op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnt.c
index cd412035e8b9a..8adddd82efe11 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_shrnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,121 +15,205 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svshrnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnt_n_s16u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svshrnt_n_s16(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svshrnt,_n_s16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnt_n_s16_1u10__SVInt8_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svshrnt_n_s16_1(svint8_t op, svint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_s16'}}
   return SVE_ACLE_FUNC(svshrnt,_n_s16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnt_n_s32u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshrnt_n_s32(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svshrnt,_n_s32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnt_n_s32_1u11__SVInt16_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svshrnt_n_s32_1(svint16_t op, svint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_s32'}}
   return SVE_ACLE_FUNC(svshrnt,_n_s32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnt_n_s64u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshrnt_n_s64(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svshrnt,_n_s64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnt_n_s64_1u11__SVInt32_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svshrnt_n_s64_1(svint32_t op, svint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_s64'}}
   return SVE_ACLE_FUNC(svshrnt,_n_s64,,)(op, op1, 32);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnt_n_u16u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svshrnt_n_u16(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svshrnt,_n_u16,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnt_n_u16_1u11__SVUint8_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> [[OP:%.*]], <vscale x 8 x i16> [[OP1:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svshrnt_n_u16_1(svuint8_t op, svuint16_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.shrnt.nxv8i16(<vscale x 16 x i8> %op, <vscale x 8 x i16> %op1, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_u16'}}
   return SVE_ACLE_FUNC(svshrnt,_n_u16,,)(op, op1, 8);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnt_n_u32u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshrnt_n_u32(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svshrnt,_n_u32,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnt_n_u32_1u12__SVUint16_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> [[OP:%.*]], <vscale x 4 x i32> [[OP1:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svshrnt_n_u32_1(svuint16_t op, svuint32_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.shrnt.nxv4i32(<vscale x 8 x i16> %op, <vscale x 4 x i32> %op1, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_u32'}}
   return SVE_ACLE_FUNC(svshrnt,_n_u32,,)(op, op1, 16);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svshrnt_n_u64u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshrnt_n_u64(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svshrnt,_n_u64,,)(op, op1, 1);
 }
 
+// CHECK-LABEL: @test_svshrnt_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svshrnt_n_u64_1u12__SVUint32_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> [[OP:%.*]], <vscale x 2 x i64> [[OP1:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svshrnt_n_u64_1(svuint32_t op, svuint64_t op1)
 {
-  // CHECK-LABEL: test_svshrnt_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.shrnt.nxv2i64(<vscale x 4 x i32> %op, <vscale x 2 x i64> %op1, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svshrnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svshrnt_n_u64'}}
   return SVE_ACLE_FUNC(svshrnt,_n_u64,,)(op, op1, 32);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sli.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sli.c
index e30fa18f5a50c..49989bdde745e 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sli.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sli.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,161 +15,273 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsli_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsli_n_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsli_n_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s8'}}
   return SVE_ACLE_FUNC(svsli,_n_s8,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsli_n_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsli_n_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 7)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s8'}}
   return SVE_ACLE_FUNC(svsli,_n_s8,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svsli_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsli_n_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsli_n_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s16'}}
   return SVE_ACLE_FUNC(svsli,_n_s16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsli_n_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsli_n_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 15)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s16'}}
   return SVE_ACLE_FUNC(svsli,_n_s16,,)(op1, op2, 15);
 }
 
+// CHECK-LABEL: @test_svsli_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsli_n_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsli_n_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s32'}}
   return SVE_ACLE_FUNC(svsli,_n_s32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsli_n_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsli_n_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 31)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s32'}}
   return SVE_ACLE_FUNC(svsli,_n_s32,,)(op1, op2, 31);
 }
 
+// CHECK-LABEL: @test_svsli_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsli_n_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsli_n_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s64'}}
   return SVE_ACLE_FUNC(svsli,_n_s64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsli_n_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsli_n_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 63)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_s64'}}
   return SVE_ACLE_FUNC(svsli,_n_s64,,)(op1, op2, 63);
 }
 
+// CHECK-LABEL: @test_svsli_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsli_n_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsli_n_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 0)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u8'}}
   return SVE_ACLE_FUNC(svsli,_n_u8,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 7)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsli_n_u8_1u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 7)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsli_n_u8_1(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sli.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 7)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u8'}}
   return SVE_ACLE_FUNC(svsli,_n_u8,,)(op1, op2, 7);
 }
 
+// CHECK-LABEL: @test_svsli_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsli_n_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsli_n_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 0)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u16'}}
   return SVE_ACLE_FUNC(svsli,_n_u16,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 15)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsli_n_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 15)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsli_n_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sli.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 15)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u16'}}
   return SVE_ACLE_FUNC(svsli,_n_u16,,)(op1, op2, 15);
 }
 
+// CHECK-LABEL: @test_svsli_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsli_n_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsli_n_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 0)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u32'}}
   return SVE_ACLE_FUNC(svsli,_n_u32,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 31)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsli_n_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 31)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsli_n_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sli.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 31)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u32'}}
   return SVE_ACLE_FUNC(svsli,_n_u32,,)(op1, op2, 31);
 }
 
+// CHECK-LABEL: @test_svsli_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsli_n_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 0)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsli_n_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 0)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u64'}}
   return SVE_ACLE_FUNC(svsli,_n_u64,,)(op1, op2, 0);
 }
 
+// CHECK-LABEL: @test_svsli_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 63)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsli_n_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 63)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsli_n_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsli_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sli.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 63)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsli'}}
   // expected-warning at +1 {{implicit declaration of function 'svsli_n_u64'}}
   return SVE_ACLE_FUNC(svsli,_n_u64,,)(op1, op2, 63);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4e.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4e.c
index b43b014cbee92..3ab3a2453d5e5 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4e.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4e.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,11 +15,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsm4e_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sm4e(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsm4e_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sm4e(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsm4e_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsm4e_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sm4e(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsm4e'}}
   // expected-warning at +1 {{implicit declaration of function 'svsm4e_u32'}}
   return SVE_ACLE_FUNC(svsm4e,_u32,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4ekey.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4ekey.c
index 488e74e3b7b01..d99aef16ee462 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4ekey.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sm4ekey.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2-sm4 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,11 +15,18 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsm4ekey_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sm4ekey(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsm4ekey_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sm4ekey(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsm4ekey_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsm4ekey_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sm4ekey(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsm4ekey'}}
   // expected-warning at +1 {{implicit declaration of function 'svsm4ekey_u32'}}
   return SVE_ACLE_FUNC(svsm4ekey,_u32,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sqadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sqadd.c
index fcb023320040e..bf162093c7ded 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sqadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sqadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 #include <arm_sve.h>
@@ -13,278 +14,487 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsqadd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqadd_u8_mu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsqadd_u8_m(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u8_m'}}
   return SVE_ACLE_FUNC(svsqadd,_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u16_mu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsqadd_u16_m(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u16_m'}}
   return SVE_ACLE_FUNC(svsqadd,_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u32_mu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsqadd_u32_m(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
   // CHECKA-LABEL: test_svsqadd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u32_m'}}
   return SVE_ACLE_FUNC(svsqadd,_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u64_mu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsqadd_u64_m(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u64_m'}}
   return SVE_ACLE_FUNC(svsqadd,_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsqadd_n_u8_mu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsqadd_n_u8_m(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u8_m'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u16_mu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsqadd_n_u16_m(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u16_m'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u32_mu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsqadd_n_u32_m(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u32_m'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u64_mu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsqadd_n_u64_m(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u64_m'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqadd_u8_zu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsqadd_u8_z(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u8_z'}}
   return SVE_ACLE_FUNC(svsqadd,_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u16_zu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsqadd_u16_z(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u16_z'}}
   return SVE_ACLE_FUNC(svsqadd,_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u32_zu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsqadd_u32_z(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u32_z'}}
   return SVE_ACLE_FUNC(svsqadd,_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u64_zu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsqadd_u64_z(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u64_z'}}
   return SVE_ACLE_FUNC(svsqadd,_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsqadd_n_u8_zu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svsqadd_n_u8_z(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u8_z
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u8_z'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u16_zu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svuint16_t test_svsqadd_n_u16_z(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u16_z'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u32_zu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svuint32_t test_svsqadd_n_u32_z(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u32_z'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u64_zu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svuint64_t test_svsqadd_n_u64_z(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u64_z'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsqadd_u8_xu10__SVBool_tu11__SVUint8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsqadd_u8_x(svbool_t pg, svuint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u8_x'}}
   return SVE_ACLE_FUNC(svsqadd,_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u16_xu10__SVBool_tu12__SVUint16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsqadd_u16_x(svbool_t pg, svuint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u16_x'}}
   return SVE_ACLE_FUNC(svsqadd,_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u32_xu10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsqadd_u32_x(svbool_t pg, svuint32_t op1, svint32_t op2)
 {
   // CHECKA-LABEL: test_svsqadd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u32_x'}}
   return SVE_ACLE_FUNC(svsqadd,_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsqadd_u64_xu10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsqadd_u64_x(svbool_t pg, svuint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_u64_x'}}
   return SVE_ACLE_FUNC(svsqadd,_u64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsqadd_n_u8_xu10__SVBool_tu11__SVUint8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsqadd_n_u8_x(svbool_t pg, svuint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u8_x'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u16_xu10__SVBool_tu12__SVUint16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svsqadd_n_u16_x(svbool_t pg, svuint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u16_x'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u32_xu10__SVBool_tu12__SVUint32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svsqadd_n_u32_x(svbool_t pg, svuint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u32_x'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svsqadd_n_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svsqadd_n_u64_xu10__SVBool_tu12__SVUint64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svsqadd_n_u64_x(svbool_t pg, svuint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsqadd_n_u64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svsqadd_n_u64_x'}}
   return SVE_ACLE_FUNC(svsqadd,_n_u64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sra.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sra.c
index f99fce73698c6..26805003303d9 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sra.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sra.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,161 +15,273 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsra_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ssra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsra_n_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ssra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsra_n_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ssra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s8'}}
   return SVE_ACLE_FUNC(svsra,_n_s8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ssra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsra_n_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ssra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsra_n_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ssra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s8'}}
   return SVE_ACLE_FUNC(svsra,_n_s8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svsra_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsra_n_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsra_n_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s16'}}
   return SVE_ACLE_FUNC(svsra,_n_s16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsra_n_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsra_n_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s16'}}
   return SVE_ACLE_FUNC(svsra,_n_s16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svsra_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsra_n_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsra_n_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s32'}}
   return SVE_ACLE_FUNC(svsra,_n_s32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsra_n_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsra_n_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s32'}}
   return SVE_ACLE_FUNC(svsra,_n_s32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svsra_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsra_n_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsra_n_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s64'}}
   return SVE_ACLE_FUNC(svsra,_n_s64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsra_n_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsra_n_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_s64'}}
   return SVE_ACLE_FUNC(svsra,_n_s64,,)(op1, op2, 64);
 }
 
+// CHECK-LABEL: @test_svsra_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsra_n_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsra_n_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u8'}}
   return SVE_ACLE_FUNC(svsra,_n_u8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsra_n_u8_1u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usra.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsra_n_u8_1(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.usra.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u8'}}
   return SVE_ACLE_FUNC(svsra,_n_u8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svsra_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsra_n_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsra_n_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u16'}}
   return SVE_ACLE_FUNC(svsra,_n_u16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsra_n_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usra.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsra_n_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usra.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u16'}}
   return SVE_ACLE_FUNC(svsra,_n_u16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svsra_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsra_n_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsra_n_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u32'}}
   return SVE_ACLE_FUNC(svsra,_n_u32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsra_n_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usra.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsra_n_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usra.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u32'}}
   return SVE_ACLE_FUNC(svsra,_n_u32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svsra_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsra_n_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsra_n_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u64'}}
   return SVE_ACLE_FUNC(svsra,_n_u64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsra_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsra_n_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usra.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsra_n_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsra_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usra.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsra'}}
   // expected-warning at +1 {{implicit declaration of function 'svsra_n_u64'}}
   return SVE_ACLE_FUNC(svsra,_n_u64,,)(op1, op2, 64);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sri.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sri.c
index 5fd74b88c3373..f6e0fe4d3885a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sri.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sri.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,161 +15,273 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsri_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsri_n_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsri_n_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s8'}}
   return SVE_ACLE_FUNC(svsri,_n_s8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsri_n_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsri_n_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s8'}}
   return SVE_ACLE_FUNC(svsri,_n_s8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svsri_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsri_n_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsri_n_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s16'}}
   return SVE_ACLE_FUNC(svsri,_n_s16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsri_n_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsri_n_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s16'}}
   return SVE_ACLE_FUNC(svsri,_n_s16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svsri_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsri_n_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsri_n_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s32'}}
   return SVE_ACLE_FUNC(svsri,_n_s32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsri_n_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsri_n_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s32'}}
   return SVE_ACLE_FUNC(svsri,_n_s32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svsri_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsri_n_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsri_n_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s64'}}
   return SVE_ACLE_FUNC(svsri,_n_s64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsri_n_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsri_n_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_s64'}}
   return SVE_ACLE_FUNC(svsri,_n_s64,,)(op1, op2, 64);
 }
 
+// CHECK-LABEL: @test_svsri_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svsri_n_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsri_n_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u8'}}
   return SVE_ACLE_FUNC(svsri,_n_u8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsri_n_u8_1u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsri_n_u8_1(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sri.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u8'}}
   return SVE_ACLE_FUNC(svsri,_n_u8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svsri_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsri_n_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsri_n_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u16'}}
   return SVE_ACLE_FUNC(svsri,_n_u16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsri_n_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsri_n_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sri.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u16'}}
   return SVE_ACLE_FUNC(svsri,_n_u16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svsri_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsri_n_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsri_n_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u32'}}
   return SVE_ACLE_FUNC(svsri,_n_u32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsri_n_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsri_n_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sri.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u32'}}
   return SVE_ACLE_FUNC(svsri,_n_u32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svsri_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsri_n_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsri_n_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u64'}}
   return SVE_ACLE_FUNC(svsri,_n_u64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svsri_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsri_n_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsri_n_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsri_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sri.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsri'}}
   // expected-warning at +1 {{implicit declaration of function 'svsri_n_u64'}}
   return SVE_ACLE_FUNC(svsri,_n_u64,,)(op1, op2, 64);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1.c
index 6d24e3f272be9..df6f6c8578c5d 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,337 +15,607 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svstnt1_scatter_u32base_s32u10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_s32(svbool_t pg, svuint32_t bases, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_s32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, , _s32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svstnt1_scatter_u64base_s64u10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_s64(svbool_t pg, svuint64_t bases, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, , _s64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svstnt1_scatter_u32base_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_u32(svbool_t pg, svuint32_t bases, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_u32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, , _u32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svstnt1_scatter_u64base_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_u64(svbool_t pg, svuint64_t bases, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, , _u64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svstnt1_scatter_u32base_f32u10__SVBool_tu12__SVUint32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_f32(svbool_t pg, svuint32_t bases, svfloat32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_f32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, , _f32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z32test_svstnt1_scatter_u64base_f64u10__SVBool_tu12__SVUint64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_f64(svbool_t pg, svuint64_t bases, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, , _f64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_s64offset_s64u10__SVBool_tPlu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_s64offset_s64(svbool_t pg, int64_t *base, svint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_s64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_s64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, s64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_s64offset_u64u10__SVBool_tPmu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_s64offset_u64(svbool_t pg, uint64_t *base, svint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_s64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_s64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, s64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_s64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_s64offset_f64u10__SVBool_tPdu11__SVInt64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_s64offset_f64(svbool_t pg, float64_t *base, svint64_t offsets, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_s64offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_s64offset_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, s64, offset, _f64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_u32offset_s32u10__SVBool_tPiu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32offset_s32(svbool_t pg, int32_t *base, svuint32_t offsets, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32offset_s32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u32, offset, _s32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_u64offset_s64u10__SVBool_tPlu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64offset_s64(svbool_t pg, int64_t *base, svuint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_u32offset_u32u10__SVBool_tPju12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], i32* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32offset_u32(svbool_t pg, uint32_t *base, svuint32_t offsets, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], i32* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32offset_u32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u32, offset, _u32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_u64offset_u64u10__SVBool_tPmu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64offset_u64(svbool_t pg, uint64_t *base, svuint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_u32offset_f32u10__SVBool_tPfu12__SVUint32_tu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4f32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], float* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32offset_f32(svbool_t pg, float32_t *base, svuint32_t offsets, svfloat32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4f32(<vscale x 4 x float> %data, <vscale x 4 x i1> [[PG]], float* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32offset_f32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u32, offset, _f32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1_scatter_u64offset_f64u10__SVBool_tPdu12__SVUint64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64offset_f64(svbool_t pg, float64_t *base, svuint64_t offsets, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64offset_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u64, offset, _f64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1_scatter_u32base_offset_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_offset_s32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, _offset, _s32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1_scatter_u64base_offset_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_offset_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, _offset, _s64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1_scatter_u32base_offset_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_offset_u32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, _offset, _u32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1_scatter_u64base_offset_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_offset_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, _offset, _u64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_offset_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1_scatter_u32base_offset_f32u10__SVBool_tu12__SVUint32_tlu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_offset_f32(svbool_t pg, svuint32_t bases, int64_t offset, svfloat32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_offset_f32
-  // CHECK: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_offset_f32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, _offset, _f32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_offset_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1_scatter_u64base_offset_f64u10__SVBool_tu12__SVUint64_tlu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_offset_f64(svbool_t pg, svuint64_t bases, int64_t offset, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_offset_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_offset_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, _offset, _f64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1_scatter_s64index_s64u10__SVBool_tPlu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_s64index_s64(svbool_t pg, int64_t *base, svint64_t indices, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_s64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_s64index_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, s64, index, _s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1_scatter_s64index_u64u10__SVBool_tPmu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_s64index_u64(svbool_t pg, uint64_t *base, svint64_t indices, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_s64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_s64index_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, s64, index, _u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_s64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1_scatter_s64index_f64u10__SVBool_tPdu11__SVInt64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_s64index_f64(svbool_t pg, float64_t *base, svint64_t indices, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_s64index_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_s64index_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, s64, index, _f64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1_scatter_u64index_s64u10__SVBool_tPlu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64index_s64(svbool_t pg, int64_t *base, svuint64_t indices, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64index_s64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64index_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u64, index, _s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1_scatter_u64index_u64u10__SVBool_tPmu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], i64* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64index_u64(svbool_t pg, uint64_t *base, svuint64_t indices, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64index_u64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], i64* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64index_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u64, index, _u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1_scatter_u64index_f64u10__SVBool_tPdu12__SVUint64_tu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2f64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], double* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64index_f64(svbool_t pg, float64_t *base, svuint64_t indices, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64index_f64
-  // CHECK: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2f64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], double* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64index_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter_, u64, index, _f64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z38test_svstnt1_scatter_u32base_index_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_index_s32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_index_s32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, _index, _s32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z38test_svstnt1_scatter_u64base_index_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_index_s64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, _index, _s64)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z38test_svstnt1_scatter_u32base_index_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_index_u32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i32.nxv4i32(<vscale x 4 x i32> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_index_u32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, _index, _u32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z38test_svstnt1_scatter_u64base_index_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_index_u64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i64.nxv2i64(<vscale x 2 x i64> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, _index, _u64)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u32base_index_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z38test_svstnt1_scatter_u32base_index_f32u10__SVBool_tu12__SVUint32_tlu13__SVFloat32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u32base_index_f32(svbool_t pg, svuint32_t bases, int64_t index, svfloat32_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u32base_index_f32
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4f32.nxv4i32(<vscale x 4 x float> %data, <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u32base_index_f32'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u32base, _index, _f32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1_scatter_u64base_index_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z38test_svstnt1_scatter_u64base_index_f64u10__SVBool_tu12__SVUint64_tlu13__SVFloat64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[INDEX:%.*]], 3
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP1]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1_scatter_u64base_index_f64(svbool_t pg, svuint64_t bases, int64_t index, svfloat64_t data) {
-  // CHECK-LABEL: test_svstnt1_scatter_u64base_index_f64
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 3
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2f64.nxv2i64(<vscale x 2 x double> %data, <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1_scatter_u64base_index_f64'}}
   return SVE_ACLE_FUNC(svstnt1_scatter, _u64base, _index, _f64)(pg, bases, index, data);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1b.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1b.c
index e8ec536916dee..d118ff9997b76 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1b.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1b.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,155 +15,281 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1b_scatter_u32base_s32u10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u32base_s32(svbool_t pg, svuint32_t bases, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u32base_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u32base_s32'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u32base, , _s32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1b_scatter_u64base_s64u10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u64base_s64(svbool_t pg, svuint64_t bases, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u64base_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u64base_s64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u64base, , _s64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1b_scatter_u32base_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u32base_u32(svbool_t pg, svuint32_t bases, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u32base_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u32base_u32'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u32base, , _u32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1b_scatter_u64base_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u64base_u64(svbool_t pg, svuint64_t bases, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u64base_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u64base_u64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u64base, , _u64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1b_scatter_s64offset_s64u10__SVBool_tPau11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_s64offset_s64(svbool_t pg, int8_t *base, svint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_s64offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_s64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter_, s64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1b_scatter_s64offset_u64u10__SVBool_tPhu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_s64offset_u64(svbool_t pg, uint8_t *base, svint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_s64offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_s64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter_, s64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i8(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1b_scatter_u32offset_s32u10__SVBool_tPau12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i8(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u32offset_s32(svbool_t pg, int8_t *base, svuint32_t offsets, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u32offset_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i8(<vscale x 4 x i8> [[TRUNC]], <vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u32offset_s32'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter_, u32, offset, _s32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1b_scatter_u64offset_s64u10__SVBool_tPau12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u64offset_s64(svbool_t pg, int8_t *base, svuint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u64offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter_, u64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i8(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1b_scatter_u32offset_u32u10__SVBool_tPhu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i8(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u32offset_u32(svbool_t pg, uint8_t *base, svuint32_t offsets, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u32offset_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i8(<vscale x 4 x i8> [[TRUNC]], <vscale x 4 x i1> [[PG]], i8* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u32offset_u32'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter_, u32, offset, _u32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1b_scatter_u64offset_u64u10__SVBool_tPhu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i8* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u64offset_u64(svbool_t pg, uint8_t *base, svuint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u64offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i8(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], i8* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter_, u64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1b_scatter_u32base_offset_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u32base_offset_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u32base, _offset, _s32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1b_scatter_u64base_offset_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u64base_offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u64base, _offset, _s64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1b_scatter_u32base_offset_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u32base_offset_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i8.nxv4i32(<vscale x 4 x i8> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u32base, _offset, _u32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1b_scatter_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1b_scatter_u64base_offset_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i8>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1b_scatter_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1b_scatter_u64base_offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i8>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i8.nxv2i64(<vscale x 2 x i8> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1b_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1b_scatter_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1b_scatter, _u64base, _offset, _u64)(pg, bases, offset, data);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1h.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1h.c
index 8463b55dcad42..498e7a9ef5b23 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1h.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1h.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,247 +15,449 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32base_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1h_scatter_u32base_s32u10__SVBool_tu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32base_s32(svbool_t pg, svuint32_t bases, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32base_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32base_s32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u32base, , _s32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1h_scatter_u64base_s64u10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64base_s64(svbool_t pg, svuint64_t bases, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64base_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64base_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u64base, , _s64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32base_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1h_scatter_u32base_u32u10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32base_u32(svbool_t pg, svuint32_t bases, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32base_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32base_u32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u32base, , _u32)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1h_scatter_u64base_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64base_u64(svbool_t pg, svuint64_t bases, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64base_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64base_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u64base, , _u64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1h_scatter_s64offset_s64u10__SVBool_tPsu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_s64offset_s64(svbool_t pg, int16_t *base, svint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_s64offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_s64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, s64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1h_scatter_s64offset_u64u10__SVBool_tPtu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_s64offset_u64(svbool_t pg, uint16_t *base, svint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_s64offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_s64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, s64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i16(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1h_scatter_u32offset_s32u10__SVBool_tPsu12__SVUint32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i16(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32offset_s32(svbool_t pg, int16_t *base, svuint32_t offsets, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32offset_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i16(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32offset_s32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, u32, offset, _s32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1h_scatter_u64offset_s64u10__SVBool_tPsu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64offset_s64(svbool_t pg, int16_t *base, svuint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, u64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i16(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1h_scatter_u32offset_u32u10__SVBool_tPtu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i16(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 4 x i32> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32offset_u32(svbool_t pg, uint16_t *base, svuint32_t offsets, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32offset_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.uxtw.nxv4i16(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], i16* %base, <vscale x 4 x i32> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32offset_u32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, u32, offset, _u32)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1h_scatter_u64offset_u64u10__SVBool_tPtu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64offset_u64(svbool_t pg, uint16_t *base, svuint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, u64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32base_offset_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1h_scatter_u32base_offset_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32base_offset_s32(svbool_t pg, svuint32_t bases, int64_t offset, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32base_offset_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32base_offset_s32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u32base, _offset, _s32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1h_scatter_u64base_offset_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64base_offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u64base, _offset, _s64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32base_offset_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1h_scatter_u32base_offset_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32base_offset_u32(svbool_t pg, svuint32_t bases, int64_t offset, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32base_offset_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32base_offset_u32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u32base, _offset, _u32)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1h_scatter_u64base_offset_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64base_offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u64base, _offset, _u64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1h_scatter_s64index_s64u10__SVBool_tPsu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_s64index_s64(svbool_t pg, int16_t *base, svint64_t indices, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_s64index_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_s64index_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, s64, index, _s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1h_scatter_s64index_u64u10__SVBool_tPtu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_s64index_u64(svbool_t pg, uint16_t *base, svint64_t indices, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_s64index_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_s64index_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, s64, index, _u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1h_scatter_u64index_s64u10__SVBool_tPsu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64index_s64(svbool_t pg, int16_t *base, svuint64_t indices, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64index_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64index_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, u64, index, _s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1h_scatter_u64index_u64u10__SVBool_tPtu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i16* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64index_u64(svbool_t pg, uint16_t *base, svuint64_t indices, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64index_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i16(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], i16* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64index_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter_, u64, index, _u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32base_index_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1h_scatter_u32base_index_s32u10__SVBool_tu12__SVUint32_tlu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32base_index_s32(svbool_t pg, svuint32_t bases, int64_t index, svint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32base_index_s32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32base_index_s32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u32base, _index, _s32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1h_scatter_u64base_index_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64base_index_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u64base, _index, _s64)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u32base_index_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1h_scatter_u32base_index_u32u10__SVBool_tu12__SVUint32_tlu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 4 x i32> [[DATA:%.*]] to <vscale x 4 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TMP0]], <vscale x 4 x i1> [[TMP1]], <vscale x 4 x i32> [[BASES:%.*]], i64 [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u32base_index_u32(svbool_t pg, svuint32_t bases, int64_t index, svuint32_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u32base_index_u32
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 4 x i32> %data to <vscale x 4 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv4i16.nxv4i32(<vscale x 4 x i16> [[TRUNC]], <vscale x 4 x i1> [[PG]], <vscale x 4 x i32> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u32base_index_u32'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u32base, _index, _u32)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1h_scatter_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1h_scatter_u64base_index_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i16>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 1
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1h_scatter_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1h_scatter_u64base_index_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i16>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 1
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i16.nxv2i64(<vscale x 2 x i16> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1h_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1h_scatter_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svstnt1h_scatter, _u64base, _index, _u64)(pg, bases, index, data);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1w.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1w.c
index 561c3e14c1a31..8ddba85d6f7dd 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1w.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_stnt1w.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,157 +15,285 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64base_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1w_scatter_u64base_s64u10__SVBool_tu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64base_s64(svbool_t pg, svuint64_t bases, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64base_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64base_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter, _u64base, , _s64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64base_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z33test_svstnt1w_scatter_u64base_u64u10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 0)
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64base_u64(svbool_t pg, svuint64_t bases, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64base_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 0)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64base_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter, _u64base, , _u64)(pg, bases, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_s64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1w_scatter_s64offset_s64u10__SVBool_tPiu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_s64offset_s64(svbool_t pg, int32_t *base, svint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_s64offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_s64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, s64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_s64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1w_scatter_s64offset_u64u10__SVBool_tPju11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_s64offset_u64(svbool_t pg, uint32_t *base, svint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_s64offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_s64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, s64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1w_scatter_u64offset_s64u10__SVBool_tPiu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64offset_s64(svbool_t pg, int32_t *base, svuint64_t offsets, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, u64, offset, _s64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z35test_svstnt1w_scatter_u64offset_u64u10__SVBool_tPju12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[OFFSETS:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64offset_u64(svbool_t pg, uint32_t *base, svuint64_t offsets, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %offsets)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, u64, offset, _u64)(pg, base, offsets, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64base_offset_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1w_scatter_u64base_offset_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64base_offset_s64(svbool_t pg, svuint64_t bases, int64_t offset, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64base_offset_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64base_offset_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter, _u64base, _offset, _s64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64base_offset_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z40test_svstnt1w_scatter_u64base_offset_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[OFFSET:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64base_offset_u64(svbool_t pg, svuint64_t bases, int64_t offset, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64base_offset_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 %offset)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_offset'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64base_offset_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter, _u64base, _offset, _u64)(pg, bases, offset, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_s64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1w_scatter_s64index_s64u10__SVBool_tPiu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_s64index_s64(svbool_t pg, int32_t *base, svint64_t indices, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_s64index_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_s64index_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, s64, index, _s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_s64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1w_scatter_s64index_u64u10__SVBool_tPju11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_s64index_u64(svbool_t pg, uint32_t *base, svint64_t indices, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_s64index_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_s64index_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, s64, index, _u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1w_scatter_u64index_s64u10__SVBool_tPiu12__SVUint64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64index_s64(svbool_t pg, int32_t *base, svuint64_t indices, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64index_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64index_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, u64, index, _s64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z34test_svstnt1w_scatter_u64index_u64u10__SVBool_tPju12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32* [[BASE:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64index_u64(svbool_t pg, uint32_t *base, svuint64_t indices, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64index_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.index.nxv2i32(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], i32* %base, <vscale x 2 x i64> %indices)
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64index_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter_, u64, index, _u64)(pg, base, indices, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64base_index_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1w_scatter_u64base_index_s64u10__SVBool_tu12__SVUint64_tlu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64base_index_s64(svbool_t pg, svuint64_t bases, int64_t index, svint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64base_index_s64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64base_index_s64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter, _u64base, _index, _s64)(pg, bases, index, data);
 }
 
+// CHECK-LABEL: @test_svstnt1w_scatter_u64base_index_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CHECK-NEXT:    ret void
+//
+// CPP-CHECK-LABEL: @_Z39test_svstnt1w_scatter_u64base_index_u64u10__SVBool_tu12__SVUint64_tlu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = trunc <vscale x 2 x i64> [[DATA:%.*]] to <vscale x 2 x i32>
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = shl i64 [[INDEX:%.*]], 2
+// CPP-CHECK-NEXT:    call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TMP0]], <vscale x 2 x i1> [[TMP1]], <vscale x 2 x i64> [[BASES:%.*]], i64 [[TMP2]])
+// CPP-CHECK-NEXT:    ret void
+//
 void test_svstnt1w_scatter_u64base_index_u64(svbool_t pg, svuint64_t bases, int64_t index, svuint64_t data) {
-  // CHECK-LABEL: test_svstnt1w_scatter_u64base_index_u64
-  // CHECK-DAG: [[TRUNC:%.*]] = trunc <vscale x 2 x i64> %data to <vscale x 2 x i32>
-  // CHECK-DAG: [[PG:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: [[SHL:%.*]] = shl i64 %index, 2
-  // CHECK: call void @llvm.aarch64.sve.stnt1.scatter.scalar.offset.nxv2i32.nxv2i64(<vscale x 2 x i32> [[TRUNC]], <vscale x 2 x i1> [[PG]], <vscale x 2 x i64> %bases, i64 [[SHL]])
-  // CHECK: ret void
   // overload-warning at +2 {{implicit declaration of function 'svstnt1w_scatter_index'}}
   // expected-warning at +1 {{implicit declaration of function 'svstnt1w_scatter_u64base_index_u64'}}
   return SVE_ACLE_FUNC(svstnt1w_scatter, _u64base, _index, _u64)(pg, bases, index, data);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnb.c
index c56d5d0f710e3..6aac7783901ee 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsubhnb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnb_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsubhnb_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_s16'}}
   return SVE_ACLE_FUNC(svsubhnb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnb_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsubhnb_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_s32'}}
   return SVE_ACLE_FUNC(svsubhnb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnb_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsubhnb_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_s64'}}
   return SVE_ACLE_FUNC(svsubhnb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnb_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsubhnb_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_u16'}}
   return SVE_ACLE_FUNC(svsubhnb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnb_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsubhnb_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_u32'}}
   return SVE_ACLE_FUNC(svsubhnb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnb_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsubhnb_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_u64'}}
   return SVE_ACLE_FUNC(svsubhnb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnb_n_s16u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsubhnb_n_s16(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_n_s16'}}
   return SVE_ACLE_FUNC(svsubhnb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnb_n_s32u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubhnb_n_s32(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_n_s32'}}
   return SVE_ACLE_FUNC(svsubhnb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnb_n_s64u11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubhnb_n_s64(svint64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_n_s64'}}
   return SVE_ACLE_FUNC(svsubhnb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnb_n_u16u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsubhnb_n_u16(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_n_u16'}}
   return SVE_ACLE_FUNC(svsubhnb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnb_n_u32u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsubhnb_n_u32(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_n_u32'}}
   return SVE_ACLE_FUNC(svsubhnb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubhnb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnb_n_u64u12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsubhnb_n_u64(svuint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svsubhnb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnb_n_u64'}}
   return SVE_ACLE_FUNC(svsubhnb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnt.c
index 4864fb9401822..8a00938b6aab5 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subhnt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsubhnt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnt_s16u10__SVInt8_tu11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svsubhnt_s16(svint8_t op1, svint16_t op2, svint16_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_s16'}}
   return SVE_ACLE_FUNC(svsubhnt,_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnt_s32u11__SVInt16_tu11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsubhnt_s32(svint16_t op1, svint32_t op2, svint32_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_s32'}}
   return SVE_ACLE_FUNC(svsubhnt,_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnt_s64u11__SVInt32_tu11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsubhnt_s64(svint32_t op1, svint64_t op2, svint64_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_s64'}}
   return SVE_ACLE_FUNC(svsubhnt,_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnt_u16u11__SVUint8_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svsubhnt_u16(svuint8_t op1, svuint16_t op2, svuint16_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %op3)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_u16'}}
   return SVE_ACLE_FUNC(svsubhnt,_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnt_u32u12__SVUint16_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsubhnt_u32(svuint16_t op1, svuint32_t op2, svuint32_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %op3)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_u32'}}
   return SVE_ACLE_FUNC(svsubhnt,_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubhnt_u64u12__SVUint32_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP3:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsubhnt_u64(svuint32_t op1, svuint64_t op2, svuint64_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %op3)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_u64'}}
   return SVE_ACLE_FUNC(svsubhnt,_u64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnt_n_s16u10__SVInt8_tu11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svsubhnt_n_s16(svint8_t op1, svint16_t op2, int16_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_n_s16'}}
   return SVE_ACLE_FUNC(svsubhnt,_n_s16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnt_n_s32u11__SVInt16_tu11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubhnt_n_s32(svint16_t op1, svint32_t op2, int32_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_n_s32'}}
   return SVE_ACLE_FUNC(svsubhnt,_n_s32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnt_n_s64u11__SVInt32_tu11__SVInt64_tl(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubhnt_n_s64(svint32_t op1, svint64_t op2, int64_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_n_s64'}}
   return SVE_ACLE_FUNC(svsubhnt,_n_s64,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnt_n_u16u11__SVUint8_tu12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svuint8_t test_svsubhnt_n_u16(svuint8_t op1, svuint16_t op2, uint16_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.subhnt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 8 x i16> %op2, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_n_u16'}}
   return SVE_ACLE_FUNC(svsubhnt,_n_u16,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnt_n_u32u12__SVUint16_tu12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsubhnt_n_u32(svuint16_t op1, svuint32_t op2, uint32_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.subhnt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 4 x i32> %op2, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_n_u32'}}
   return SVE_ACLE_FUNC(svsubhnt,_n_u32,,)(op1, op2, op3);
 }
 
+// CHECK-LABEL: @test_svsubhnt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubhnt_n_u64u12__SVUint32_tu12__SVUint64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP3:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsubhnt_n_u64(svuint32_t op1, svuint64_t op2, uint64_t op3)
 {
-  // CHECK-LABEL: test_svsubhnt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op3)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.subhnt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 2 x i64> %op2, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubhnt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubhnt_n_u64'}}
   return SVE_ACLE_FUNC(svsubhnt,_n_u64,,)(op1, op2, op3);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublb.c
index 988be977ce2da..93845fcbf7604 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsublb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublb_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsublb_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsublb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_s16'}}
   return SVE_ACLE_FUNC(svsublb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublb_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsublb_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsublb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_s32'}}
   return SVE_ACLE_FUNC(svsublb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublb_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsublb_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsublb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_s64'}}
   return SVE_ACLE_FUNC(svsublb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublb_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsublb_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsublb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_u16'}}
   return SVE_ACLE_FUNC(svsublb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublb_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsublb_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsublb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_u32'}}
   return SVE_ACLE_FUNC(svsublb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublb_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsublb_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsublb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_u64'}}
   return SVE_ACLE_FUNC(svsublb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublb_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsublb_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsublb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_n_s16'}}
   return SVE_ACLE_FUNC(svsublb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublb_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsublb_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsublb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_n_s32'}}
   return SVE_ACLE_FUNC(svsublb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublb_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsublb_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsublb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_n_s64'}}
   return SVE_ACLE_FUNC(svsublb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublb_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsublb_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsublb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_n_u16'}}
   return SVE_ACLE_FUNC(svsublb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublb_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsublb_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsublb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_n_u32'}}
   return SVE_ACLE_FUNC(svsublb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublb_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsublb_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsublb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublb_n_u64'}}
   return SVE_ACLE_FUNC(svsublb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublbt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublbt.c
index 02afe3e0cbcb4..674b2f7c64b5a 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublbt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublbt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,64 +15,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsublbt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsublbt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsublbt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsublbt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublbt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublbt_s16'}}
   return SVE_ACLE_FUNC(svsublbt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublbt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsublbt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsublbt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsublbt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublbt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublbt_s32'}}
   return SVE_ACLE_FUNC(svsublbt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublbt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsublbt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsublbt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsublbt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublbt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublbt_s64'}}
   return SVE_ACLE_FUNC(svsublbt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublbt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsublbt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublbt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsublbt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsublbt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublbt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublbt_n_s16'}}
   return SVE_ACLE_FUNC(svsublbt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublbt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsublbt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublbt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsublbt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsublbt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublbt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublbt_n_s32'}}
   return SVE_ACLE_FUNC(svsublbt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublbt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsublbt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublbt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsublbt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsublbt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublbt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublbt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublbt_n_s64'}}
   return SVE_ACLE_FUNC(svsublbt,_n_s64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublt.c
index 4a8bd9635043b..25742e28ff143 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_sublt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsublt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublt_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsublt_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsublt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_s16'}}
   return SVE_ACLE_FUNC(svsublt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublt_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsublt_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsublt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_s32'}}
   return SVE_ACLE_FUNC(svsublt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublt_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsublt_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsublt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_s64'}}
   return SVE_ACLE_FUNC(svsublt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublt_u16u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsublt_u16(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsublt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_u16'}}
   return SVE_ACLE_FUNC(svsublt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublt_u32u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsublt_u32(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsublt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_u32'}}
   return SVE_ACLE_FUNC(svsublt,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsublt_u64u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsublt_u64(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsublt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_u64'}}
   return SVE_ACLE_FUNC(svsublt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublt_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsublt_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsublt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssublt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_n_s16'}}
   return SVE_ACLE_FUNC(svsublt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublt_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsublt_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsublt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssublt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_n_s32'}}
   return SVE_ACLE_FUNC(svsublt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublt_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsublt_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsublt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssublt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_n_s64'}}
   return SVE_ACLE_FUNC(svsublt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublt_n_u16u11__SVUint8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublt.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsublt_n_u16(svuint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsublt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usublt.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_n_u16'}}
   return SVE_ACLE_FUNC(svsublt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublt_n_u32u12__SVUint16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublt.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsublt_n_u32(svuint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsublt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usublt.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_n_u32'}}
   return SVE_ACLE_FUNC(svsublt,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsublt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsublt_n_u64u12__SVUint32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublt.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsublt_n_u64(svuint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsublt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usublt.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsublt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsublt_n_u64'}}
   return SVE_ACLE_FUNC(svsublt,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subltb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subltb.c
index 36c7561145661..89cb511ca6684 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subltb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subltb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,64 +15,109 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsubltb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubltb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubltb_s16u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubltb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsubltb_s16(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsubltb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubltb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubltb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubltb_s16'}}
   return SVE_ACLE_FUNC(svsubltb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubltb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubltb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubltb_s32u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubltb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsubltb_s32(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubltb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubltb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubltb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubltb_s32'}}
   return SVE_ACLE_FUNC(svsubltb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubltb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubltb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svsubltb_s64u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubltb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsubltb_s64(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubltb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubltb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubltb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubltb_s64'}}
   return SVE_ACLE_FUNC(svsubltb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubltb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubltb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubltb_n_s16u10__SVInt8_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubltb.nxv8i16(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubltb_n_s16(svint8_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsubltb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubltb.nxv8i16(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubltb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubltb_n_s16'}}
   return SVE_ACLE_FUNC(svsubltb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubltb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubltb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubltb_n_s32u11__SVInt16_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubltb.nxv4i32(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubltb_n_s32(svint16_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubltb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubltb.nxv4i32(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubltb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubltb_n_s32'}}
   return SVE_ACLE_FUNC(svsubltb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubltb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubltb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svsubltb_n_s64u11__SVInt32_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubltb.nxv2i64(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsubltb_n_s64(svint32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubltb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubltb.nxv2i64(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubltb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubltb_n_s64'}}
   return SVE_ACLE_FUNC(svsubltb,_n_s64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwb.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwb.c
index 1f86b5edb252e..b3519e4110b2b 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwb.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwb.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsubwb_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwb_s16u11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsubwb_s16(svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_s16'}}
   return SVE_ACLE_FUNC(svsubwb,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwb_s32u11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsubwb_s32(svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_s32'}}
   return SVE_ACLE_FUNC(svsubwb,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwb_s64u11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsubwb_s64(svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_s64'}}
   return SVE_ACLE_FUNC(svsubwb,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwb_u16u12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsubwb_u16(svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_u16'}}
   return SVE_ACLE_FUNC(svsubwb,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwb_u32u12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsubwb_u32(svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_u32'}}
   return SVE_ACLE_FUNC(svsubwb,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwb_u64u12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsubwb_u64(svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_u64'}}
   return SVE_ACLE_FUNC(svsubwb,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwb_n_s16u11__SVInt16_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubwb_n_s16(svint16_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_n_s16'}}
   return SVE_ACLE_FUNC(svsubwb,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwb_n_s32u11__SVInt32_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubwb_n_s32(svint32_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_n_s32'}}
   return SVE_ACLE_FUNC(svsubwb,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwb_n_s64u11__SVInt64_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsubwb_n_s64(svint64_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_n_s64'}}
   return SVE_ACLE_FUNC(svsubwb,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwb_n_u16u12__SVUint16_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwb.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsubwb_n_u16(svuint16_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwb.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_n_u16'}}
   return SVE_ACLE_FUNC(svsubwb,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwb_n_u32u12__SVUint32_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwb.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsubwb_n_u32(svuint32_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwb.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_n_u32'}}
   return SVE_ACLE_FUNC(svsubwb,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwb_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwb_n_u64u12__SVUint64_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwb.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsubwb_n_u64(svuint64_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsubwb_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwb.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwb'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwb_n_u64'}}
   return SVE_ACLE_FUNC(svsubwb,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwt.c
index 9e72444ce958c..3b2a2da7710ce 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_subwt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,127 +15,217 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svsubwt_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwt_s16u11__SVInt16_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svsubwt_s16(svint16_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_s16'}}
   return SVE_ACLE_FUNC(svsubwt,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwt_s32u11__SVInt32_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svsubwt_s32(svint32_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_s32'}}
   return SVE_ACLE_FUNC(svsubwt,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwt_s64u11__SVInt64_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svsubwt_s64(svint64_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_s64'}}
   return SVE_ACLE_FUNC(svsubwt,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwt_u16u12__SVUint16_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svsubwt_u16(svuint16_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_u16'}}
   return SVE_ACLE_FUNC(svsubwt,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwt_u32u12__SVUint32_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svsubwt_u32(svuint32_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_u32'}}
   return SVE_ACLE_FUNC(svsubwt,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svsubwt_u64u12__SVUint64_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svsubwt_u64(svuint64_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_u64'}}
   return SVE_ACLE_FUNC(svsubwt,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwt_n_s16u11__SVInt16_ta(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svsubwt_n_s16(svint16_t op1, int8_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_n_s16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ssubwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_n_s16'}}
   return SVE_ACLE_FUNC(svsubwt,_n_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwt_n_s32u11__SVInt32_ts(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svsubwt_n_s32(svint32_t op1, int16_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_n_s32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ssubwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_n_s32'}}
   return SVE_ACLE_FUNC(svsubwt,_n_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwt_n_s64u11__SVInt64_ti(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svsubwt_n_s64(svint64_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_n_s64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ssubwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_n_s64'}}
   return SVE_ACLE_FUNC(svsubwt,_n_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwt_n_u16u12__SVUint16_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwt.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svuint16_t test_svsubwt_n_u16(svuint16_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_n_u16
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.usubwt.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_n_u16'}}
   return SVE_ACLE_FUNC(svsubwt,_n_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwt_n_u32u12__SVUint32_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwt.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 8 x i16> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svuint32_t test_svsubwt_n_u32(svuint32_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_n_u32
-  // CHECK: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.usubwt.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_n_u32'}}
   return SVE_ACLE_FUNC(svsubwt,_n_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svsubwt_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svsubwt_n_u64u12__SVUint64_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwt.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 4 x i32> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svuint64_t test_svsubwt_n_u64(svuint64_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svsubwt_n_u64
-  // CHECK: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.usubwt.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svsubwt'}}
   // expected-warning at +1 {{implicit declaration of function 'svsubwt_n_u64'}}
   return SVE_ACLE_FUNC(svsubwt,_n_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
index 0e3434c15f22f..db4ad45a31770 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2-bfloat.c
@@ -1,8 +1,9 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +bf16 -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
@@ -16,12 +17,21 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtbl2_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl2.nxv8bf16(<vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svtbl2_bf1614svbfloat16x2_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl2.nxv8bf16(<vscale x 8 x bfloat> [[TMP0]], <vscale x 8 x bfloat> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP2]]
+//
 svbfloat16_t test_svtbl2_bf16(svbfloat16x2_t data, svuint16_t indices) {
-  // CHECK-LABEL: test_svtbl2_bf16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tuple.get.nxv8bf16.nxv16bf16(<vscale x 16 x bfloat> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbl2.nxv8bf16(<vscale x 8 x bfloat> %[[V0]], <vscale x 8 x bfloat> %[[V1]], <vscale x 8 x i16> %indices)
-  // CHECK-NEXT: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_bf16'}}
   return SVE_ACLE_FUNC(svtbl2, _bf16, , )(data, indices);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
index 936ffedd00860..920de4c286508 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbl2.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,133 +15,232 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtbl2_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl2_s810svint8x2_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svtbl2_s8(svint8x2_t data, svuint8_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_s8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %indices)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_s8'}}
   return SVE_ACLE_FUNC(svtbl2,_s8,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_s1611svint16x2_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svtbl2_s16(svint16x2_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_s16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %indices)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_s16'}}
   return SVE_ACLE_FUNC(svtbl2,_s16,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_s3211svint32x2_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svtbl2_s32(svint32x2_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_s32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %indices)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_s32'}}
   return SVE_ACLE_FUNC(svtbl2,_s32,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_s6411svint64x2_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svtbl2_s64(svint64x2_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_s64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %indices)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_s64'}}
   return SVE_ACLE_FUNC(svtbl2,_s64,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbl2_u811svuint8x2_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svuint8_t test_svtbl2_u8(svuint8x2_t data, svuint8_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_u8
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tuple.get.nxv16i8.nxv32i8(<vscale x 32 x i8> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbl2.nxv16i8(<vscale x 16 x i8> %[[V0]], <vscale x 16 x i8> %[[V1]], <vscale x 16 x i8> %indices)
-  // CHECK-NEXT: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_u8'}}
   return SVE_ACLE_FUNC(svtbl2,_u8,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_u1612svuint16x2_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svuint16_t test_svtbl2_u16(svuint16x2_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_u16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tuple.get.nxv8i16.nxv16i16(<vscale x 16 x i16> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbl2.nxv8i16(<vscale x 8 x i16> %[[V0]], <vscale x 8 x i16> %[[V1]], <vscale x 8 x i16> %indices)
-  // CHECK-NEXT: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_u16'}}
   return SVE_ACLE_FUNC(svtbl2,_u16,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_u3212svuint32x2_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svuint32_t test_svtbl2_u32(svuint32x2_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_u32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tuple.get.nxv4i32.nxv8i32(<vscale x 8 x i32> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbl2.nxv4i32(<vscale x 4 x i32> %[[V0]], <vscale x 4 x i32> %[[V1]], <vscale x 4 x i32> %indices)
-  // CHECK-NEXT: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_u32'}}
   return SVE_ACLE_FUNC(svtbl2,_u32,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_u6412svuint64x2_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svuint64_t test_svtbl2_u64(svuint64x2_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_u64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tuple.get.nxv2i64.nxv4i64(<vscale x 4 x i64> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbl2.nxv2i64(<vscale x 2 x i64> %[[V0]], <vscale x 2 x i64> %[[V1]], <vscale x 2 x i64> %indices)
-  // CHECK-NEXT: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_u64'}}
   return SVE_ACLE_FUNC(svtbl2,_u64,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl2.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_f1613svfloat16x2_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl2.nxv8f16(<vscale x 8 x half> [[TMP0]], <vscale x 8 x half> [[TMP1]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP2]]
+//
 svfloat16_t test_svtbl2_f16(svfloat16x2_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_f16
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tuple.get.nxv8f16.nxv16f16(<vscale x 16 x half> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbl2.nxv8f16(<vscale x 8 x half> %[[V0]], <vscale x 8 x half> %[[V1]], <vscale x 8 x i16> %indices)
-  // CHECK-NEXT: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_f16'}}
   return SVE_ACLE_FUNC(svtbl2,_f16,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl2.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_f3213svfloat32x2_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl2.nxv4f32(<vscale x 4 x float> [[TMP0]], <vscale x 4 x float> [[TMP1]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP2]]
+//
 svfloat32_t test_svtbl2_f32(svfloat32x2_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_f32
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tuple.get.nxv4f32.nxv8f32(<vscale x 8 x float> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbl2.nxv4f32(<vscale x 4 x float> %[[V0]], <vscale x 4 x float> %[[V1]], <vscale x 4 x i32> %indices)
-  // CHECK-NEXT: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_f32'}}
   return SVE_ACLE_FUNC(svtbl2,_f32,,)(data, indices);
 }
 
+// CHECK-LABEL: @test_svtbl2_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl2.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbl2_f6413svfloat64x2_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA:%.*]], i32 0)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> [[DATA]], i32 1)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl2.nxv2f64(<vscale x 2 x double> [[TMP0]], <vscale x 2 x double> [[TMP1]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP2]]
+//
 svfloat64_t test_svtbl2_f64(svfloat64x2_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbl2_f64
-  // CHECK-DAG: %[[V0:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %data, i32 0)
-  // CHECK-DAG: %[[V1:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tuple.get.nxv2f64.nxv4f64(<vscale x 4 x double> %data, i32 1)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbl2.nxv2f64(<vscale x 2 x double> %[[V0]], <vscale x 2 x double> %[[V1]], <vscale x 2 x i64> %indices)
-  // CHECK-NEXT: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbl2'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbl2_f64'}}
   return SVE_ACLE_FUNC(svtbl2,_f64,,)(data, indices);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx-bfloat.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx-bfloat.c
index 127910279b1df..480e7e4273091 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx-bfloat.c
@@ -1,8 +1,9 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // REQUIRES: aarch64-registered-target
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +bf16 -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
@@ -16,10 +17,17 @@
 #define SVE_ACLE_FUNC(A1, A2, A3, A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtbx_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbx.nxv8bf16(<vscale x 8 x bfloat> [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svtbx_bf16u14__SVBFloat16_tu14__SVBFloat16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbx.nxv8bf16(<vscale x 8 x bfloat> [[FALLBACK:%.*]], <vscale x 8 x bfloat> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
 svbfloat16_t test_svtbx_bf16(svbfloat16_t fallback, svbfloat16_t data, svuint16_t indices) {
-  // CHECK-LABEL: test_svtbx_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.tbx.nxv8bf16(<vscale x 8 x bfloat> %fallback, <vscale x 8 x bfloat> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_bf16'}}
   return SVE_ACLE_FUNC(svtbx, _bf16, , )(fallback, data, indices);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx.c
index c31918dbfaafc..61ad99478f771 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_tbx.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,111 +15,188 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svtbx_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbx.nxv16i8(<vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svtbx_s8u10__SVInt8_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbx.nxv16i8(<vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svtbx_s8(svint8_t fallback, svint8_t data, svuint8_t indices)
 {
-  // CHECK-LABEL: test_svtbx_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbx.nxv16i8(<vscale x 16 x i8> %fallback, <vscale x 16 x i8> %data, <vscale x 16 x i8> %indices)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_s8'}}
   return SVE_ACLE_FUNC(svtbx,_s8,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbx.nxv8i16(<vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_s16u11__SVInt16_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbx.nxv8i16(<vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svtbx_s16(svint16_t fallback, svint16_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbx_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbx.nxv8i16(<vscale x 8 x i16> %fallback, <vscale x 8 x i16> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_s16'}}
   return SVE_ACLE_FUNC(svtbx,_s16,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbx.nxv4i32(<vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_s32u11__SVInt32_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbx.nxv4i32(<vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svtbx_s32(svint32_t fallback, svint32_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbx_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbx.nxv4i32(<vscale x 4 x i32> %fallback, <vscale x 4 x i32> %data, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_s32'}}
   return SVE_ACLE_FUNC(svtbx,_s32,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbx.nxv2i64(<vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_s64u11__SVInt64_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbx.nxv2i64(<vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svtbx_s64(svint64_t fallback, svint64_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbx_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbx.nxv2i64(<vscale x 2 x i64> %fallback, <vscale x 2 x i64> %data, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_s64'}}
   return SVE_ACLE_FUNC(svtbx,_s64,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbx.nxv16i8(<vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z13test_svtbx_u8u11__SVUint8_tu11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbx.nxv16i8(<vscale x 16 x i8> [[FALLBACK:%.*]], <vscale x 16 x i8> [[DATA:%.*]], <vscale x 16 x i8> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svtbx_u8(svuint8_t fallback, svuint8_t data, svuint8_t indices)
 {
-  // CHECK-LABEL: test_svtbx_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.tbx.nxv16i8(<vscale x 16 x i8> %fallback, <vscale x 16 x i8> %data, <vscale x 16 x i8> %indices)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_u8'}}
   return SVE_ACLE_FUNC(svtbx,_u8,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbx.nxv8i16(<vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_u16u12__SVUint16_tu12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbx.nxv8i16(<vscale x 8 x i16> [[FALLBACK:%.*]], <vscale x 8 x i16> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svtbx_u16(svuint16_t fallback, svuint16_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbx_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.tbx.nxv8i16(<vscale x 8 x i16> %fallback, <vscale x 8 x i16> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_u16'}}
   return SVE_ACLE_FUNC(svtbx,_u16,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbx.nxv4i32(<vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_u32u12__SVUint32_tu12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbx.nxv4i32(<vscale x 4 x i32> [[FALLBACK:%.*]], <vscale x 4 x i32> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svtbx_u32(svuint32_t fallback, svuint32_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbx_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.tbx.nxv4i32(<vscale x 4 x i32> %fallback, <vscale x 4 x i32> %data, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_u32'}}
   return SVE_ACLE_FUNC(svtbx,_u32,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbx.nxv2i64(<vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_u64u12__SVUint64_tu12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbx.nxv2i64(<vscale x 2 x i64> [[FALLBACK:%.*]], <vscale x 2 x i64> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svtbx_u64(svuint64_t fallback, svuint64_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbx_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.tbx.nxv2i64(<vscale x 2 x i64> %fallback, <vscale x 2 x i64> %data, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_u64'}}
   return SVE_ACLE_FUNC(svtbx,_u64,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbx.nxv8f16(<vscale x 8 x half> [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_f16u13__SVFloat16_tu13__SVFloat16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbx.nxv8f16(<vscale x 8 x half> [[FALLBACK:%.*]], <vscale x 8 x half> [[DATA:%.*]], <vscale x 8 x i16> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x half> [[TMP0]]
+//
 svfloat16_t test_svtbx_f16(svfloat16_t fallback, svfloat16_t data, svuint16_t indices)
 {
-  // CHECK-LABEL: test_svtbx_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.tbx.nxv8f16(<vscale x 8 x half> %fallback, <vscale x 8 x half> %data, <vscale x 8 x i16> %indices)
-  // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_f16'}}
   return SVE_ACLE_FUNC(svtbx,_f16,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbx.nxv4f32(<vscale x 4 x float> [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_f32u13__SVFloat32_tu13__SVFloat32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbx.nxv4f32(<vscale x 4 x float> [[FALLBACK:%.*]], <vscale x 4 x float> [[DATA:%.*]], <vscale x 4 x i32> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
 svfloat32_t test_svtbx_f32(svfloat32_t fallback, svfloat32_t data, svuint32_t indices)
 {
-  // CHECK-LABEL: test_svtbx_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.tbx.nxv4f32(<vscale x 4 x float> %fallback, <vscale x 4 x float> %data, <vscale x 4 x i32> %indices)
-  // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_f32'}}
   return SVE_ACLE_FUNC(svtbx,_f32,,)(fallback, data, indices);
 }
 
+// CHECK-LABEL: @test_svtbx_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbx.nxv2f64(<vscale x 2 x double> [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z14test_svtbx_f64u13__SVFloat64_tu13__SVFloat64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbx.nxv2f64(<vscale x 2 x double> [[FALLBACK:%.*]], <vscale x 2 x double> [[DATA:%.*]], <vscale x 2 x i64> [[INDICES:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x double> [[TMP0]]
+//
 svfloat64_t test_svtbx_f64(svfloat64_t fallback, svfloat64_t data, svuint64_t indices)
 {
-  // CHECK-LABEL: test_svtbx_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.tbx.nxv2f64(<vscale x 2 x double> %fallback, <vscale x 2 x double> %data, <vscale x 2 x i64> %indices)
-  // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svtbx'}}
   // expected-warning at +1 {{implicit declaration of function 'svtbx_f64'}}
   return SVE_ACLE_FUNC(svtbx,_f64,,)(fallback, data, indices);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_uqadd.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_uqadd.c
index 72b96cfc84531..35fc04b811914 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_uqadd.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_uqadd.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 #include <arm_sve.h>
@@ -13,278 +14,487 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svuqadd_u8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svuqadd_u8_mu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svuqadd_u8_m(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u8_m
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s8_m'}}
   return SVE_ACLE_FUNC(svuqadd,_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u16_mu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svuqadd_u16_m(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u16_m
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s16_m'}}
   return SVE_ACLE_FUNC(svuqadd,_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u32_mu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svuqadd_u32_m(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svuqadd_u32_m
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s32_m'}}
   return SVE_ACLE_FUNC(svuqadd,_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u64_mu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svuqadd_u64_m(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u64_m
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s64_m'}}
   return SVE_ACLE_FUNC(svuqadd,_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s8_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svuqadd_n_s8_mu10__SVBool_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svuqadd_n_s8_m(svbool_t pg, svint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s8_m
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s8_m'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s8,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s16_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s16_mu10__SVBool_tu11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svuqadd_n_s16_m(svbool_t pg, svint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s16_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s16_m'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s16,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s32_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s32_mu10__SVBool_tu11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svuqadd_n_s32_m(svbool_t pg, svint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s32_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s32_m'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s32,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s64_m(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s64_mu10__SVBool_tu11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svuqadd_n_s64_m(svbool_t pg, svint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s64_m
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_m'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s64_m'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s64,_m,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svuqadd_u8_zu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP0]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svuqadd_u8_z(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u8_z
-  // CHECK: %[[SEL:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s8_z'}}
   return SVE_ACLE_FUNC(svuqadd,_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u16_zu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP1]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svuqadd_u16_z(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u16_z
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s16_z'}}
   return SVE_ACLE_FUNC(svuqadd,_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u32_zu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP1]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svuqadd_u32_z(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u32_z
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s32_z'}}
   return SVE_ACLE_FUNC(svuqadd,_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u64_zu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP1]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svuqadd_u64_z(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u64_z
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s64_z'}}
   return SVE_ACLE_FUNC(svuqadd,_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s8_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svuqadd_n_s8_zu10__SVBool_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.sel.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG]], <vscale x 16 x i8> [[TMP1]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP2]]
+//
 svint8_t test_svuqadd_n_s8_z(svbool_t pg, svint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s8_z
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %[[SEL]], <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s8_z'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s8,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s16_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s16_zu10__SVBool_tu11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[TMP2]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP3]]
+//
 svint16_t test_svuqadd_n_s16_z(svbool_t pg, svint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s16_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.sel.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %[[SEL]], <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s16_z'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s16,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s32_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s32_zu10__SVBool_tu11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[TMP2]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP3]]
+//
 svint32_t test_svuqadd_n_s32_z(svbool_t pg, svint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s32_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.sel.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %[[SEL]], <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s32_z'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s32,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s64_z(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s64_zu10__SVBool_tu11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> zeroinitializer)
+// CPP-CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[TMP2]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP3]]
+//
 svint64_t test_svuqadd_n_s64_z(svbool_t pg, svint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s64_z
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK-DAG: %[[SEL:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.sel.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> zeroinitializer)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %[[SEL]], <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_z'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s64_z'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s64,_z,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svuqadd_u8_xu10__SVBool_tu10__SVInt8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svuqadd_u8_x(svbool_t pg, svint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u8_x
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s8_x'}}
   return SVE_ACLE_FUNC(svuqadd,_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u16_xu10__SVBool_tu11__SVInt16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
+//
 svint16_t test_svuqadd_u16_x(svbool_t pg, svint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u16_x
-  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s16_x'}}
   return SVE_ACLE_FUNC(svuqadd,_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u32_xu10__SVBool_tu11__SVInt32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
+//
 svint32_t test_svuqadd_u32_x(svbool_t pg, svint32_t op1, svuint32_t op2)
 {
   // CHECKA-LABEL: test_svuqadd_u32_x
-  // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s32_x'}}
   return SVE_ACLE_FUNC(svuqadd,_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_u64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svuqadd_u64_xu10__SVBool_tu11__SVInt64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
+//
 svint64_t test_svuqadd_u64_x(svbool_t pg, svint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_u64_x
-  // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_s64_x'}}
   return SVE_ACLE_FUNC(svuqadd,_s64,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s8_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svuqadd_n_s8_xu10__SVBool_tu10__SVInt8_th(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP1]]
+//
 svint8_t test_svuqadd_n_s8_x(svbool_t pg, svint8_t op1, uint8_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s8_x
-  // CHECK: %[[DUP:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.dup.x.nxv16i8(i8 %op2)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.suqadd.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %op1, <vscale x 16 x i8> %[[DUP]])
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s8_x'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s8,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s16_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s16_xu10__SVBool_tu11__SVInt16_tt(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP2]]
+//
 svint16_t test_svuqadd_n_s16_x(svbool_t pg, svint16_t op1, uint16_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s16_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.dup.x.nxv8i16(i16 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.suqadd.nxv8i16(<vscale x 8 x i1> %[[PG]], <vscale x 8 x i16> %op1, <vscale x 8 x i16> %[[DUP]])
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s16_x'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s16,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s32_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s32_xu10__SVBool_tu11__SVInt32_tj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP2]]
+//
 svint32_t test_svuqadd_n_s32_x(svbool_t pg, svint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s32_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.dup.x.nxv4i32(i32 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.suqadd.nxv4i32(<vscale x 4 x i1> %[[PG]], <vscale x 4 x i32> %op1, <vscale x 4 x i32> %[[DUP]])
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s32_x'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s32,_x,)(pg, op1, op2);
 }
 
+// CHECK-LABEL: @test_svuqadd_n_s64_x(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svuqadd_n_s64_xu10__SVBool_tu11__SVInt64_tm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[TMP1]])
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP2]]
+//
 svint64_t test_svuqadd_n_s64_x(svbool_t pg, svint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svuqadd_n_s64_x
-  // CHECK-DAG: %[[DUP:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.dup.x.nxv2i64(i64 %op2)
-  // CHECK-DAG: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.suqadd.nxv2i64(<vscale x 2 x i1> %[[PG]], <vscale x 2 x i64> %op1, <vscale x 2 x i64> %[[DUP]])
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svuqadd_x'}}
   // expected-warning at +1 {{implicit declaration of function 'svuqadd_n_s64_x'}}
   return SVE_ACLE_FUNC(svuqadd,_n_s64,_x,)(pg, op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilege.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilege.c
index ac37dbdfc0b15..d9d1897f8910f 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilege.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilege.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,173 +15,297 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilege_b8_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilege.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilege_b8_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilege.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilege_b8_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b8_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilege.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b8_s32'}}
   return SVE_ACLE_FUNC(svwhilege_b8,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilege.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b16_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilege.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b16_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b16_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilege.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b16_s32'}}
   return SVE_ACLE_FUNC(svwhilege_b16,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilege.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b32_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilege.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b32_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b32_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilege.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b32_s32'}}
   return SVE_ACLE_FUNC(svwhilege_b32,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilege.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b64_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilege.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b64_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b64_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilege.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b64_s32'}}
   return SVE_ACLE_FUNC(svwhilege_b64,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b8_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehs.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilege_b8_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehs.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilege_b8_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b8_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehs.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b8_u32'}}
   return SVE_ACLE_FUNC(svwhilege_b8,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehs.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b16_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehs.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b16_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b16_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehs.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b16_u32'}}
   return SVE_ACLE_FUNC(svwhilege_b16,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehs.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b32_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehs.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b32_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b32_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehs.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b32_u32'}}
   return SVE_ACLE_FUNC(svwhilege_b32,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehs.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b64_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehs.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b64_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b64_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehs.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b64_u32'}}
   return SVE_ACLE_FUNC(svwhilege_b64,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b8_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilege.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilege_b8_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilege.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilege_b8_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b8_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilege.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b8_s64'}}
   return SVE_ACLE_FUNC(svwhilege_b8,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilege.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b16_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilege.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b16_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b16_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilege.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b16_s64'}}
   return SVE_ACLE_FUNC(svwhilege_b16,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilege.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b32_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilege.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b32_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b32_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilege.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b32_s64'}}
   return SVE_ACLE_FUNC(svwhilege_b32,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilege.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b64_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilege.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b64_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b64_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilege.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b64_s64'}}
   return SVE_ACLE_FUNC(svwhilege_b64,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b8_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehs.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilege_b8_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehs.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilege_b8_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b8_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehs.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b8_u64'}}
   return SVE_ACLE_FUNC(svwhilege_b8,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehs.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b16_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehs.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b16_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b16_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehs.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b16_u64'}}
   return SVE_ACLE_FUNC(svwhilege_b16,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehs.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b32_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehs.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b32_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b32_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehs.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b32_u64'}}
   return SVE_ACLE_FUNC(svwhilege_b32,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilege_b64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehs.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilege_b64_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehs.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilege_b64_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilege_b64_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehs.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilege_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilege_b64_u64'}}
   return SVE_ACLE_FUNC(svwhilege_b64,_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilegt.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilegt.c
index 0186ad1f75846..0c7d3bb062ab4 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilegt.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilegt.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,173 +15,297 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilegt_b8_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilegt.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilegt_b8_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilegt.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilegt_b8_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b8_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilegt.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b8_s32'}}
   return SVE_ACLE_FUNC(svwhilegt_b8,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b16_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilegt.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b16_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilegt.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b16_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b16_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilegt.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b16_s32'}}
   return SVE_ACLE_FUNC(svwhilegt_b16,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b32_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilegt.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b32_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilegt.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b32_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b32_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilegt.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b32_s32'}}
   return SVE_ACLE_FUNC(svwhilegt_b32,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b64_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilegt.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b64_s32ii(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilegt.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b64_s32(int32_t op1, int32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b64_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilegt.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b64_s32'}}
   return SVE_ACLE_FUNC(svwhilegt_b64,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b8_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehi.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilegt_b8_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehi.nxv16i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilegt_b8_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b8_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehi.nxv16i1.i32(i32 %op1, i32 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b8_u32'}}
   return SVE_ACLE_FUNC(svwhilegt_b8,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b16_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehi.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b16_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehi.nxv8i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b16_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b16_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehi.nxv8i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b16_u32'}}
   return SVE_ACLE_FUNC(svwhilegt_b16,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b32_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehi.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b32_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehi.nxv4i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b32_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b32_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehi.nxv4i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b32_u32'}}
   return SVE_ACLE_FUNC(svwhilegt_b32,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b64_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehi.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b64_u32jj(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehi.nxv2i1.i32(i32 [[OP1:%.*]], i32 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b64_u32(uint32_t op1, uint32_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b64_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehi.nxv2i1.i32(i32 %op1, i32 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b64_u32'}}
   return SVE_ACLE_FUNC(svwhilegt_b64,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b8_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilegt.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilegt_b8_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilegt.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilegt_b8_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b8_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilegt.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b8_s64'}}
   return SVE_ACLE_FUNC(svwhilegt_b8,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b16_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilegt.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b16_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilegt.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b16_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b16_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilegt.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b16_s64'}}
   return SVE_ACLE_FUNC(svwhilegt_b16,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b32_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilegt.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b32_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilegt.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b32_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b32_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilegt.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b32_s64'}}
   return SVE_ACLE_FUNC(svwhilegt_b32,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b64_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilegt.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b64_s64ll(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilegt.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b64_s64(int64_t op1, int64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b64_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilegt.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b64_s64'}}
   return SVE_ACLE_FUNC(svwhilegt_b64,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b8_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehi.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z21test_svwhilegt_b8_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehi.nxv16i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilegt_b8_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b8_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilehi.nxv16i1.i64(i64 %op1, i64 %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b8'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b8_u64'}}
   return SVE_ACLE_FUNC(svwhilegt_b8,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b16_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehi.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b16_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehi.nxv8i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b16_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b16_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilehi.nxv8i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b16'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b16_u64'}}
   return SVE_ACLE_FUNC(svwhilegt_b16,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b32_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehi.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b32_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehi.nxv4i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b32_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b32_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilehi.nxv4i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b32'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b32_u64'}}
   return SVE_ACLE_FUNC(svwhilegt_b32,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilegt_b64_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehi.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z22test_svwhilegt_b64_u64mm(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehi.nxv2i1.i64(i64 [[OP1:%.*]], i64 [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilegt_b64_u64(uint64_t op1, uint64_t op2)
 {
-  // CHECK-LABEL: test_svwhilegt_b64_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilehi.nxv2i1.i64(i64 %op1, i64 %op2)
-  // CHECK: %[[CAST:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[CAST]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilegt_b64'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilegt_b64_u64'}}
   return SVE_ACLE_FUNC(svwhilegt_b64,_u64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw-bfloat.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw-bfloat.c
index bebd40a0dbd33..7ad27588ee7e4 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,12 +15,20 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilerw_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0bf16(bfloat* [[OP1:%.*]], bfloat* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svwhilerw_bf16PKu6__bf16S0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0bf16(bfloat* [[OP1:%.*]], bfloat* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_bf16(const bfloat16_t *op1, const bfloat16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0bf16(bfloat* %op1, bfloat* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_bf16'}}
   return SVE_ACLE_FUNC(svwhilerw,_bf16,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw.c
index 1bd486d746279..9aa97f2a39830 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilerw.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,119 +15,206 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilerw_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilerw.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svwhilerw_s8PKaS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilerw.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilerw_s8(const int8_t *op1, const int8_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilerw.b.nxv16i1.p0i8(i8* %op1, i8* %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_s8'}}
   return SVE_ACLE_FUNC(svwhilerw,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_s16PKsS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_s16(const int16_t *op1, const int16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0i16(i16* %op1, i16* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_s16'}}
   return SVE_ACLE_FUNC(svwhilerw,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_s32PKiS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_s32(const int32_t *op1, const int32_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0i32(i32* %op1, i32* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_s32'}}
   return SVE_ACLE_FUNC(svwhilerw,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_s64PKlS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_s64(const int64_t *op1, const int64_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0i64(i64* %op1, i64* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_s64'}}
   return SVE_ACLE_FUNC(svwhilerw,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilerw.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svwhilerw_u8PKhS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilerw.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilerw_u8(const uint8_t *op1, const uint8_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilerw.b.nxv16i1.p0i8(i8* %op1, i8* %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_u8'}}
   return SVE_ACLE_FUNC(svwhilerw,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_u16PKtS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_u16(const uint16_t *op1, const uint16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0i16(i16* %op1, i16* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_u16'}}
   return SVE_ACLE_FUNC(svwhilerw,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_u32PKjS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_u32(const uint32_t *op1, const uint32_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0i32(i32* %op1, i32* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_u32'}}
   return SVE_ACLE_FUNC(svwhilerw,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_u64PKmS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_u64(const uint64_t *op1, const uint64_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0i64(i64* %op1, i64* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_u64'}}
   return SVE_ACLE_FUNC(svwhilerw,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0f16(half* [[OP1:%.*]], half* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_f16PKDhS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0f16(half* [[OP1:%.*]], half* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_f16(const float16_t *op1, const float16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilerw.h.nxv8i1.p0f16(half* %op1, half* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_f16'}}
   return SVE_ACLE_FUNC(svwhilerw,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0f32(float* [[OP1:%.*]], float* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_f32PKfS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0f32(float* [[OP1:%.*]], float* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_f32(const float32_t *op1, const float32_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilerw.s.nxv4i1.p0f32(float* %op1, float* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_f32'}}
   return SVE_ACLE_FUNC(svwhilerw,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilerw_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0f64(double* [[OP1:%.*]], double* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilerw_f64PKdS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0f64(double* [[OP1:%.*]], double* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilerw_f64(const float64_t *op1, const float64_t *op2)
 {
-  // CHECK-LABEL: test_svwhilerw_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilerw.d.nxv2i1.p0f64(double* %op1, double* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
   // overload-warning at +2 {{implicit declaration of function 'svwhilerw'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilerw_f64'}}
   return SVE_ACLE_FUNC(svwhilerw,_f64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr-bfloat.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr-bfloat.c
index c2ed031ae08bf..0a9702610a4a8 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr-bfloat.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr-bfloat.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_SVE2 -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,12 +15,20 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilewr_bf16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0bf16(bfloat* [[OP1:%.*]], bfloat* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svwhilewr_bf16PKu6__bf16S0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0bf16(bfloat* [[OP1:%.*]], bfloat* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_bf16(const bfloat16_t *op1, const bfloat16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_bf16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0bf16(bfloat* %op1, bfloat* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_bf16'}}
   return SVE_ACLE_FUNC(svwhilewr,_bf16,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr.c
index 84f4c15606083..d58eb127b8b7d 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_whilewr.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,119 +15,206 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svwhilewr_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilewr.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svwhilewr_s8PKaS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilewr.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilewr_s8(const int8_t *op1, const int8_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilewr.b.nxv16i1.p0i8(i8* %op1, i8* %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_s8'}}
   return SVE_ACLE_FUNC(svwhilewr,_s8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_s16PKsS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_s16(const int16_t *op1, const int16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0i16(i16* %op1, i16* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_s16'}}
   return SVE_ACLE_FUNC(svwhilewr,_s16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_s32PKiS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_s32(const int32_t *op1, const int32_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0i32(i32* %op1, i32* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_s32'}}
   return SVE_ACLE_FUNC(svwhilewr,_s32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_s64PKlS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_s64(const int64_t *op1, const int64_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0i64(i64* %op1, i64* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_s64'}}
   return SVE_ACLE_FUNC(svwhilewr,_s64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilewr.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svwhilewr_u8PKhS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilewr.b.nxv16i1.p0i8(i8* [[OP1:%.*]], i8* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
 svbool_t test_svwhilewr_u8(const uint8_t *op1, const uint8_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.whilewr.b.nxv16i1.p0i8(i8* %op1, i8* %op2)
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_u8'}}
   return SVE_ACLE_FUNC(svwhilewr,_u8,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_u16PKtS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0i16(i16* [[OP1:%.*]], i16* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_u16(const uint16_t *op1, const uint16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0i16(i16* %op1, i16* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_u16'}}
   return SVE_ACLE_FUNC(svwhilewr,_u16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_u32PKjS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0i32(i32* [[OP1:%.*]], i32* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_u32(const uint32_t *op1, const uint32_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0i32(i32* %op1, i32* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_u32'}}
   return SVE_ACLE_FUNC(svwhilewr,_u32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_u64PKmS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0i64(i64* [[OP1:%.*]], i64* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_u64(const uint64_t *op1, const uint64_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0i64(i64* %op1, i64* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_u64'}}
   return SVE_ACLE_FUNC(svwhilewr,_u64,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_f16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0f16(half* [[OP1:%.*]], half* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_f16PKDhS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0f16(half* [[OP1:%.*]], half* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_f16(const float16_t *op1, const float16_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_f16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.whilewr.h.nxv8i1.p0f16(half* %op1, half* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_f16'}}
   return SVE_ACLE_FUNC(svwhilewr,_f16,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_f32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0f32(float* [[OP1:%.*]], float* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_f32PKfS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0f32(float* [[OP1:%.*]], float* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_f32(const float32_t *op1, const float32_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_f32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.whilewr.s.nxv4i1.p0f32(float* %op1, float* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %[[INTRINSIC]])
-  // CHECK: ret <vscale x 16 x i1> %[[INTRINSIC_REINT]]
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_f32'}}
   return SVE_ACLE_FUNC(svwhilewr,_f32,,)(op1, op2);
 }
 
+// CHECK-LABEL: @test_svwhilewr_f64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0f64(double* [[OP1:%.*]], double* [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svwhilewr_f64PKdS0_(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0f64(double* [[OP1:%.*]], double* [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP0]])
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i1> [[TMP1]]
+//
 svbool_t test_svwhilewr_f64(const float64_t *op1, const float64_t *op2)
 {
-  // CHECK-LABEL: test_svwhilewr_f64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.whilewr.d.nxv2i1.p0f64(double* %op1, double* %op2)
-  // CHECK: %[[INTRINSIC_REINT:.*]] = call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %[[INTRINSIC]])
   // overload-warning at +2 {{implicit declaration of function 'svwhilewr'}}
   // expected-warning at +1 {{implicit declaration of function 'svwhilewr_f64'}}
   return SVE_ACLE_FUNC(svwhilewr,_f64,,)(op1, op2);

diff  --git a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_xar.c b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_xar.c
index e235f37e386ac..534752b870b71 100644
--- a/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_xar.c
+++ b/clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_xar.c
@@ -1,7 +1,8 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s
+// RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve2 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error %s
 // RUN: %clang_cc1 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -fsyntax-only -verify=overload -verify-ignore-unexpected=error %s
 
@@ -14,161 +15,273 @@
 #define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
 #endif
 
+// CHECK-LABEL: @test_svxar_n_s8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svxar_n_s8u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svxar_n_s8(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s8'}}
   return SVE_ACLE_FUNC(svxar,_n_s8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_s8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svxar_n_s8_1u10__SVInt8_tu10__SVInt8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svint8_t test_svxar_n_s8_1(svint8_t op1, svint8_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s8'}}
   return SVE_ACLE_FUNC(svxar,_n_s8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svxar_n_s16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svxar_n_s16u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svxar_n_s16(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s16'}}
   return SVE_ACLE_FUNC(svxar,_n_s16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_s16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svxar_n_s16_1u11__SVInt16_tu11__SVInt16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svint16_t test_svxar_n_s16_1(svint16_t op1, svint16_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s16'}}
   return SVE_ACLE_FUNC(svxar,_n_s16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svxar_n_s32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svxar_n_s32u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svxar_n_s32(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s32'}}
   return SVE_ACLE_FUNC(svxar,_n_s32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_s32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svxar_n_s32_1u11__SVInt32_tu11__SVInt32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svint32_t test_svxar_n_s32_1(svint32_t op1, svint32_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s32'}}
   return SVE_ACLE_FUNC(svxar,_n_s32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svxar_n_s64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svxar_n_s64u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svxar_n_s64(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s64'}}
   return SVE_ACLE_FUNC(svxar,_n_s64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_s64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svxar_n_s64_1u11__SVInt64_tu11__SVInt64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svint64_t test_svxar_n_s64_1(svint64_t op1, svint64_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_s64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_s64'}}
   return SVE_ACLE_FUNC(svxar,_n_s64,,)(op1, op2, 64);
 }
 
+// CHECK-LABEL: @test_svxar_n_u8(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z15test_svxar_n_u8u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svxar_n_u8(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u8
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 1)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u8'}}
   return SVE_ACLE_FUNC(svxar,_n_u8,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_u8_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z17test_svxar_n_u8_1u11__SVUint8_tu11__SVUint8_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i32 8)
+// CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
 svuint8_t test_svxar_n_u8_1(svuint8_t op1, svuint8_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u8_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.xar.nxv16i8(<vscale x 16 x i8> %op1, <vscale x 16 x i8> %op2, i32 8)
-  // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u8'}}
   return SVE_ACLE_FUNC(svxar,_n_u8,,)(op1, op2, 8);
 }
 
+// CHECK-LABEL: @test_svxar_n_u16(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svxar_n_u16u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svxar_n_u16(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u16
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 1)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u16'}}
   return SVE_ACLE_FUNC(svxar,_n_u16,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_u16_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svxar_n_u16_1u12__SVUint16_tu12__SVUint16_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i32 16)
+// CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
 svuint16_t test_svxar_n_u16_1(svuint16_t op1, svuint16_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u16_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.xar.nxv8i16(<vscale x 8 x i16> %op1, <vscale x 8 x i16> %op2, i32 16)
-  // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u16'}}
   return SVE_ACLE_FUNC(svxar,_n_u16,,)(op1, op2, 16);
 }
 
+// CHECK-LABEL: @test_svxar_n_u32(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svxar_n_u32u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svxar_n_u32(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u32
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 1)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u32'}}
   return SVE_ACLE_FUNC(svxar,_n_u32,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_u32_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svxar_n_u32_1u12__SVUint32_tu12__SVUint32_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i32 32)
+// CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
 svuint32_t test_svxar_n_u32_1(svuint32_t op1, svuint32_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u32_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.xar.nxv4i32(<vscale x 4 x i32> %op1, <vscale x 4 x i32> %op2, i32 32)
-  // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u32'}}
   return SVE_ACLE_FUNC(svxar,_n_u32,,)(op1, op2, 32);
 }
 
+// CHECK-LABEL: @test_svxar_n_u64(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z16test_svxar_n_u64u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 1)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svxar_n_u64(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u64
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 1)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u64'}}
   return SVE_ACLE_FUNC(svxar,_n_u64,,)(op1, op2, 1);
 }
 
+// CHECK-LABEL: @test_svxar_n_u64_1(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z18test_svxar_n_u64_1u12__SVUint64_tu12__SVUint64_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i32 64)
+// CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
 svuint64_t test_svxar_n_u64_1(svuint64_t op1, svuint64_t op2)
 {
-  // CHECK-LABEL: test_svxar_n_u64_1
-  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.xar.nxv2i64(<vscale x 2 x i64> %op1, <vscale x 2 x i64> %op2, i32 64)
-  // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
   // overload-warning at +2 {{implicit declaration of function 'svxar'}}
   // expected-warning at +1 {{implicit declaration of function 'svxar_n_u64'}}
   return SVE_ACLE_FUNC(svxar,_n_u64,,)(op1, op2, 64);


        


More information about the cfe-commits mailing list